From nobody Sat Oct 4 08:08:45 2025 Received: from mail-ua1-f73.google.com (mail-ua1-f73.google.com [209.85.222.73]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 8F75E13C3F6 for ; Mon, 18 Aug 2025 19:05:17 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.222.73 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543919; cv=none; b=jGgRwb8B4Hz+GsLkz5CiZ3k7rz1TLm8oqBwvyizBtFkjPctfqfOZZ7B8vQfH0IEnPyLkZammCPmBpaFLcFKleKGAVSTLRxIC7tFNXE+W9TJ2xtOm6l70kBZDHGNG9bd06ZY3bZ3luxJ5S+WDa2QrupklUuQ3vEnuz5gFRtd4h8w= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543919; c=relaxed/simple; bh=qa4R15qOtzR/BH3eD0r26mPAJuG5D1h7By2gS57JXe4=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=emCv2Maz5gVzN7Z+Dd53Yd6ej5eNEeeoULGLdUOPf+0bNR3uqPJybXM1zKNZkWdLeG9Yu2Tc3qZ2s3iUvUqsWQldTWGlAC2d+wwgkT3yqjnwu5kfVKT7Ut/iHKH2paxHa9z1a8caMFrSnu90wnpHaUc63dv6gjtJcGfEfuLPEA0= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=fBkhZOTm; arc=none smtp.client-ip=209.85.222.73 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="fBkhZOTm" Received: by mail-ua1-f73.google.com with SMTP id a1e0cc1a2514c-89018f681bbso10672700241.1 for ; Mon, 18 Aug 2025 12:05:17 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543916; x=1756148716; darn=vger.kernel.org; h=to:from:subject:message-id:references:mime-version:in-reply-to:date :from:to:cc:subject:date:message-id:reply-to; bh=AeQy6U6Mcs+KaMWERQdrKYPgECEKs1fDnJTxQRr5IIg=; b=fBkhZOTmDW1Qq9cScgjCtBJKDryRuCkpfYdmwjXk1uR6i0tI8ica+GK+WKnGJZvA9O DNfnNBWvcJOEyJT+M2PG2PJE+7wHPr2/H/tbyMv/bon36exYVuel8Mw41jPYlRl2jBts IFLFIzsWWfcUiWcvRA5yIYZVqbt/aKJNEnJ/KtotS79GjNXBO6EyM2H09yf91QgerZaJ x87YYf7ZkZlLwmnE4lfsvbTMR9s47BsjNHxPGobcd1RO2AGrDshiKVw9UrZrxDH6z2WA IMenRNMf7JvQbTKgAjf3rQntuDK4Q6EF6+93QbBhcsN/UM87J933lnoeDfWLDecDHEua GuTA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543916; x=1756148716; h=to:from:subject:message-id:references:mime-version:in-reply-to:date :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=AeQy6U6Mcs+KaMWERQdrKYPgECEKs1fDnJTxQRr5IIg=; b=fTgqGN8/HcbFAaEDS++rFR2T6sXWYbrakxVyBb85iGGwNrtEb9GFBtjF1NKKFaNOMe 5YZjTq1HLqXKrkJzPbZoob7UbS5lwwFzhXKGoYWMcr9LD4rQD46tkuvP5jxzvrZ8yQI6 XGjARRm9mmz6FX44UZfXtB8kGxNpDn/OuQ79DgmgEpxyEF67hSuTQb7XMWm/vYyM/EMu 9U2feQWJscCDyoWDk2yT4T3m0eGNV21Yx1WNxl3/YRqQYbWhBafiN0bEf4cF9OCFfsFH TpO3S05kl3aWq8Eupa3XABAx3ARSjZ9BNOFtnDe+gPHbsGYG14YiSNXzoNmPpk3ZnuD8 h9yA== X-Forwarded-Encrypted: i=1; AJvYcCXPvF0dk7gvvyk4HKzUnqcyM8TS9IdQRZlK/eNpQ7A1ORKviPPF/ldSi2ZlE/L2/YP85ISbuuCAE7ek8hw=@vger.kernel.org X-Gm-Message-State: AOJu0Yyxpyf/b4oi/BAwPFXBPPPUBDrdFQgbVB7QvoPGkJzKLYiTO4pX ZgboN2ZBVvJXajTIXHvzJQss3hq3ZEPNjAaHUOvI5rBmaPVbbqX05IyI7Agpr0cyE0f/Vm5h9xm 2pm60fnchwA== X-Google-Smtp-Source: AGHT+IEfHh5zGZXDJ9T3IT6pTGa5q70K/Gu2z6P2zjPHw2vIVBLhzL3OtHQqqFDhXcyKpJAh+Pcx/6lmk97i X-Received: from qkntz8.prod.google.com ([2002:a05:620a:6908:b0:7e2:e9ed:1448]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a05:6102:dcb:b0:4fa:d2c:4ff with SMTP id ada2fe7eead31-514cb1a5babmr4171144137.27.1755543916305; Mon, 18 Aug 2025 12:05:16 -0700 (PDT) Date: Mon, 18 Aug 2025 12:03:57 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-2-irogers@google.com> Subject: [PATCH v2 01/20] perf parse-events: Handle fake PMUs in CPU terms From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" The "Parsing of PMU event table metrics with fake PMUs" will test metrics on machines/models that may be missing a PMU, in such a case the fake_pmu should be used to avoid errors. Metrics that get the cpumask from a different PMU, such as "tsc/cpu=3Dcpu_atom/", also need to be resilient in this test. The parse_events_state fake_pmu is set when missing PMUs should be ignored. So that it can be queried, pass it to the config term functions, as well as to get_config_cpu, then ignore failures when fake_pmu is set. Some minor code refactoring to cut down on the indent and remove some redundant checks. Fixes: bd741d80dc65 ("perf parse-events: Allow the cpu term to be a PMU or = CPU range") Signed-off-by: Ian Rogers --- tools/perf/util/parse-events.c | 116 +++++++++++++++++---------------- 1 file changed, 60 insertions(+), 56 deletions(-) diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 8282ddf68b98..0026cff4d69e 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -126,7 +126,8 @@ static char *get_config_name(const struct parse_events_= terms *head_terms) return get_config_str(head_terms, PARSE_EVENTS__TERM_TYPE_NAME); } =20 -static struct perf_cpu_map *get_config_cpu(const struct parse_events_terms= *head_terms) +static struct perf_cpu_map *get_config_cpu(const struct parse_events_terms= *head_terms, + bool fake_pmu) { struct parse_events_term *term; struct perf_cpu_map *cpus =3D NULL; @@ -135,24 +136,33 @@ static struct perf_cpu_map *get_config_cpu(const stru= ct parse_events_terms *head return NULL; =20 list_for_each_entry(term, &head_terms->terms, list) { - if (term->type_term =3D=3D PARSE_EVENTS__TERM_TYPE_CPU) { - struct perf_cpu_map *term_cpus; + struct perf_cpu_map *term_cpus; =20 - if (term->type_val =3D=3D PARSE_EVENTS__TERM_TYPE_NUM) { - term_cpus =3D perf_cpu_map__new_int(term->val.num); + if (term->type_term !=3D PARSE_EVENTS__TERM_TYPE_CPU) + continue; + + if (term->type_val =3D=3D PARSE_EVENTS__TERM_TYPE_NUM) { + term_cpus =3D perf_cpu_map__new_int(term->val.num); + } else { + struct perf_pmu *pmu =3D perf_pmus__find(term->val.str); + + if (pmu) { + term_cpus =3D pmu->is_core && perf_cpu_map__is_empty(pmu->cpus) + ? cpu_map__online() + : perf_cpu_map__get(pmu->cpus); } else { - struct perf_pmu *pmu =3D perf_pmus__find(term->val.str); - - if (pmu && perf_cpu_map__is_empty(pmu->cpus)) - term_cpus =3D pmu->is_core ? cpu_map__online() : NULL; - else if (pmu) - term_cpus =3D perf_cpu_map__get(pmu->cpus); - else - term_cpus =3D perf_cpu_map__new(term->val.str); + term_cpus =3D perf_cpu_map__new(term->val.str); + if (!term_cpus && fake_pmu) { + /* + * Assume the PMU string makes sense on a different + * machine and fake a value with all online CPUs. + */ + term_cpus =3D cpu_map__online(); + } } - perf_cpu_map__merge(&cpus, term_cpus); - perf_cpu_map__put(term_cpus); } + perf_cpu_map__merge(&cpus, term_cpus); + perf_cpu_map__put(term_cpus); } =20 return cpus; @@ -369,13 +379,13 @@ static int parse_aliases(const char *str, const char = *const names[][EVSEL__MAX_A =20 typedef int config_term_func_t(struct perf_event_attr *attr, struct parse_events_term *term, - struct parse_events_error *err); + struct parse_events_state *parse_state); static int config_term_common(struct perf_event_attr *attr, struct parse_events_term *term, - struct parse_events_error *err); + struct parse_events_state *parse_state); static int config_attr(struct perf_event_attr *attr, const struct parse_events_terms *head, - struct parse_events_error *err, + struct parse_events_state *parse_state, config_term_func_t config_term); =20 /** @@ -471,7 +481,7 @@ int parse_events_add_cache(struct list_head *list, int = *idx, const char *name, bool found_supported =3D false; const char *config_name =3D get_config_name(parsed_terms); const char *metric_id =3D get_config_metric_id(parsed_terms); - struct perf_cpu_map *cpus =3D get_config_cpu(parsed_terms); + struct perf_cpu_map *cpus =3D get_config_cpu(parsed_terms, parse_state->f= ake_pmu); int ret =3D 0; struct evsel *first_wildcard_match =3D NULL; =20 @@ -514,8 +524,7 @@ int parse_events_add_cache(struct list_head *list, int = *idx, const char *name, found_supported =3D true; =20 if (parsed_terms) { - if (config_attr(&attr, parsed_terms, parse_state->error, - config_term_common)) { + if (config_attr(&attr, parsed_terms, parse_state, config_term_common)) { ret =3D -EINVAL; goto out_err; } @@ -767,8 +776,7 @@ int parse_events_add_breakpoint(struct parse_events_sta= te *parse_state, attr.sample_period =3D 1; =20 if (head_config) { - if (config_attr(&attr, head_config, parse_state->error, - config_term_common)) + if (config_attr(&attr, head_config, parse_state, config_term_common)) return -EINVAL; =20 if (get_config_terms(head_config, &config_terms)) @@ -903,12 +911,12 @@ void parse_events__shrink_config_terms(void) =20 static int config_term_common(struct perf_event_attr *attr, struct parse_events_term *term, - struct parse_events_error *err) + struct parse_events_state *parse_state) { -#define CHECK_TYPE_VAL(type) \ -do { \ - if (check_type_val(term, err, PARSE_EVENTS__TERM_TYPE_ ## type)) \ - return -EINVAL; \ +#define CHECK_TYPE_VAL(type) \ +do { \ + if (check_type_val(term, parse_state->error, PARSE_EVENTS__TERM_TYPE_ ## = type)) \ + return -EINVAL; \ } while (0) =20 switch (term->type_term) { @@ -939,7 +947,7 @@ do { \ if (strcmp(term->val.str, "no") && parse_branch_str(term->val.str, &attr->branch_sample_type)) { - parse_events_error__handle(err, term->err_val, + parse_events_error__handle(parse_state->error, term->err_val, strdup("invalid branch sample type"), NULL); return -EINVAL; @@ -948,7 +956,7 @@ do { \ case PARSE_EVENTS__TERM_TYPE_TIME: CHECK_TYPE_VAL(NUM); if (term->val.num > 1) { - parse_events_error__handle(err, term->err_val, + parse_events_error__handle(parse_state->error, term->err_val, strdup("expected 0 or 1"), NULL); return -EINVAL; @@ -990,7 +998,7 @@ do { \ case PARSE_EVENTS__TERM_TYPE_PERCORE: CHECK_TYPE_VAL(NUM); if ((unsigned int)term->val.num > 1) { - parse_events_error__handle(err, term->err_val, + parse_events_error__handle(parse_state->error, term->err_val, strdup("expected 0 or 1"), NULL); return -EINVAL; @@ -1005,7 +1013,7 @@ do { \ case PARSE_EVENTS__TERM_TYPE_AUX_SAMPLE_SIZE: CHECK_TYPE_VAL(NUM); if (term->val.num > UINT_MAX) { - parse_events_error__handle(err, term->err_val, + parse_events_error__handle(parse_state->error, term->err_val, strdup("too big"), NULL); return -EINVAL; @@ -1016,7 +1024,7 @@ do { \ =20 if (term->type_val =3D=3D PARSE_EVENTS__TERM_TYPE_NUM) { if (term->val.num >=3D (u64)cpu__max_present_cpu().cpu) { - parse_events_error__handle(err, term->err_val, + parse_events_error__handle(parse_state->error, term->err_val, strdup("too big"), /*help=3D*/NULL); return -EINVAL; @@ -1028,8 +1036,8 @@ do { \ break; =20 map =3D perf_cpu_map__new(term->val.str); - if (!map) { - parse_events_error__handle(err, term->err_val, + if (!map && !parse_state->fake_pmu) { + parse_events_error__handle(parse_state->error, term->err_val, strdup("not a valid PMU or CPU number"), /*help=3D*/NULL); return -EINVAL; @@ -1042,7 +1050,7 @@ do { \ case PARSE_EVENTS__TERM_TYPE_LEGACY_CACHE: case PARSE_EVENTS__TERM_TYPE_HARDWARE: default: - parse_events_error__handle(err, term->err_term, + parse_events_error__handle(parse_state->error, term->err_term, strdup(parse_events__term_type_str(term->type_term)), parse_events_formats_error_string(NULL)); return -EINVAL; @@ -1057,7 +1065,7 @@ do { \ * if an invalid config term is provided for legacy events * (for example, instructions/badterm/...), which is confusing. */ - if (!config_term_avail(term->type_term, err)) + if (!config_term_avail(term->type_term, parse_state->error)) return -EINVAL; return 0; #undef CHECK_TYPE_VAL @@ -1065,7 +1073,7 @@ do { \ =20 static int config_term_pmu(struct perf_event_attr *attr, struct parse_events_term *term, - struct parse_events_error *err) + struct parse_events_state *parse_state) { if (term->type_term =3D=3D PARSE_EVENTS__TERM_TYPE_LEGACY_CACHE) { struct perf_pmu *pmu =3D perf_pmus__find_by_type(attr->type); @@ -1074,7 +1082,7 @@ static int config_term_pmu(struct perf_event_attr *at= tr, char *err_str; =20 if (asprintf(&err_str, "Failed to find PMU for type %d", attr->type) >= =3D 0) - parse_events_error__handle(err, term->err_term, + parse_events_error__handle(parse_state->error, term->err_term, err_str, /*help=3D*/NULL); return -EINVAL; } @@ -1100,7 +1108,7 @@ static int config_term_pmu(struct perf_event_attr *at= tr, char *err_str; =20 if (asprintf(&err_str, "Failed to find PMU for type %d", attr->type) >= =3D 0) - parse_events_error__handle(err, term->err_term, + parse_events_error__handle(parse_state->error, term->err_term, err_str, /*help=3D*/NULL); return -EINVAL; } @@ -1128,12 +1136,12 @@ static int config_term_pmu(struct perf_event_attr *= attr, */ return 0; } - return config_term_common(attr, term, err); + return config_term_common(attr, term, parse_state); } =20 static int config_term_tracepoint(struct perf_event_attr *attr, struct parse_events_term *term, - struct parse_events_error *err) + struct parse_events_state *parse_state) { switch (term->type_term) { case PARSE_EVENTS__TERM_TYPE_CALLGRAPH: @@ -1147,7 +1155,7 @@ static int config_term_tracepoint(struct perf_event_a= ttr *attr, case PARSE_EVENTS__TERM_TYPE_AUX_OUTPUT: case PARSE_EVENTS__TERM_TYPE_AUX_ACTION: case PARSE_EVENTS__TERM_TYPE_AUX_SAMPLE_SIZE: - return config_term_common(attr, term, err); + return config_term_common(attr, term, parse_state); case PARSE_EVENTS__TERM_TYPE_USER: case PARSE_EVENTS__TERM_TYPE_CONFIG: case PARSE_EVENTS__TERM_TYPE_CONFIG1: @@ -1166,12 +1174,10 @@ static int config_term_tracepoint(struct perf_event= _attr *attr, case PARSE_EVENTS__TERM_TYPE_HARDWARE: case PARSE_EVENTS__TERM_TYPE_CPU: default: - if (err) { - parse_events_error__handle(err, term->err_term, + parse_events_error__handle(parse_state->error, term->err_term, strdup(parse_events__term_type_str(term->type_term)), strdup("valid terms: call-graph,stack-size\n") ); - } return -EINVAL; } =20 @@ -1180,13 +1186,13 @@ static int config_term_tracepoint(struct perf_event= _attr *attr, =20 static int config_attr(struct perf_event_attr *attr, const struct parse_events_terms *head, - struct parse_events_error *err, + struct parse_events_state *parse_state, config_term_func_t config_term) { struct parse_events_term *term; =20 list_for_each_entry(term, &head->terms, list) - if (config_term(attr, term, err)) + if (config_term(attr, term, parse_state)) return -EINVAL; =20 return 0; @@ -1378,8 +1384,7 @@ int parse_events_add_tracepoint(struct parse_events_s= tate *parse_state, if (head_config) { struct perf_event_attr attr; =20 - if (config_attr(&attr, head_config, err, - config_term_tracepoint)) + if (config_attr(&attr, head_config, parse_state, config_term_tracepoint)) return -EINVAL; } =20 @@ -1408,8 +1413,7 @@ static int __parse_events_add_numeric(struct parse_ev= ents_state *parse_state, } =20 if (head_config) { - if (config_attr(&attr, head_config, parse_state->error, - config_term_common)) + if (config_attr(&attr, head_config, parse_state, config_term_common)) return -EINVAL; =20 if (get_config_terms(head_config, &config_terms)) @@ -1418,7 +1422,7 @@ static int __parse_events_add_numeric(struct parse_ev= ents_state *parse_state, =20 name =3D get_config_name(head_config); metric_id =3D get_config_metric_id(head_config); - cpus =3D get_config_cpu(head_config); + cpus =3D get_config_cpu(head_config, parse_state->fake_pmu); ret =3D __add_event(list, &parse_state->idx, &attr, /*init_attr*/true, na= me, metric_id, pmu, &config_terms, first_wildcard_match, cpus, /*alternate_hw_config=3D*/PERF_COUNT_HW_MAX) ? 0 : -ENOMEM; @@ -1531,7 +1535,7 @@ static int parse_events_add_pmu(struct parse_events_s= tate *parse_state, fix_raw(&parsed_terms, pmu); =20 /* Configure attr/terms with a known PMU, this will set hardcoded terms. = */ - if (config_attr(&attr, &parsed_terms, parse_state->error, config_term_pmu= )) { + if (config_attr(&attr, &parsed_terms, parse_state, config_term_pmu)) { parse_events_terms__exit(&parsed_terms); return -EINVAL; } @@ -1555,7 +1559,7 @@ static int parse_events_add_pmu(struct parse_events_s= tate *parse_state, =20 /* Configure attr/terms again if an alias was expanded. */ if (alias_rewrote_terms && - config_attr(&attr, &parsed_terms, parse_state->error, config_term_pmu= )) { + config_attr(&attr, &parsed_terms, parse_state, config_term_pmu)) { parse_events_terms__exit(&parsed_terms); return -EINVAL; } @@ -1583,7 +1587,7 @@ static int parse_events_add_pmu(struct parse_events_s= tate *parse_state, return -EINVAL; } =20 - term_cpu =3D get_config_cpu(&parsed_terms); + term_cpu =3D get_config_cpu(&parsed_terms, parse_state->fake_pmu); evsel =3D __add_event(list, &parse_state->idx, &attr, /*init_attr=3D*/tru= e, get_config_name(&parsed_terms), get_config_metric_id(&parsed_terms), pmu, --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pl1-f202.google.com (mail-pl1-f202.google.com [209.85.214.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id F08E5274676 for ; Mon, 18 Aug 2025 19:05:21 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.202 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543930; cv=none; b=BZN6LuGHjgenovHw9/GmVKMlRTT0tKWZQvS1MXS9NHhXT2Bx1TNV2dTdPex4c8LldKwJoHs7yWGwDLOrz2kM9Wcir0d/HnUhOlgPhNzhveWZnuYM9JLqDr0stPdCRSdaB9TfDXQoDWK1scGpXEjDdyVeMre0H35XmgCArGT9z/o= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543930; c=relaxed/simple; bh=VFb1CaJ0eRxh9IO9QxjPvXJk8l4UbGZXT8gl5bmUsvA=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=haqUQ1zZ9kvIonrSTq+zrwqmdXekFr8wYKIqHN9XUEDmWEV4Zg/3zVbZTcjD0V424+Ow0J+Nfsb487cVpNqzmW7NUAM7GKEWcyM0hp4vatcZqhOCKntapp23lUrFWdoG0eaSUjjIunpB7URjKdgUOXmpAWet/BUBJ2CrNBMYuNE= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=E0QSk1ZL; arc=none smtp.client-ip=209.85.214.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="E0QSk1ZL" Received: by mail-pl1-f202.google.com with SMTP id d9443c01a7336-244581187c6so49083255ad.1 for ; Mon, 18 Aug 2025 12:05:21 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543921; x=1756148721; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=XPu7HgEqjv/9pYy7ZOBzTh5eusrrnsbsNUri9J4cP3I=; b=E0QSk1ZLVOkkEITpT1R3rOivICs61fsiajbZzwU4deSaJLRly6Ek8KAU2xLKQIZtDq JqJ0XypYmgK2ZlVUw8685UnZy2yZnGrvkVHRDfTRCgU5e57O7wD+scLMOO89dN30lP8r rnSx7+mi0MBohIZYzf7K+jgXgsVCfFDekquHy7NzqakDPMDJjGyaAlyUbr6+vEQVHAzD blX9ciKcUxXNOxZKHo1UInjjPQ7ZxHCcsCl+RakcKzHGAaAsfGuGjOjHcKoFNffnhplF wPnOOjp6g65xc4DQ5OlO5V5rwE3xuzyBF1ELQYbvV9kPl1Dk4IMAOvbUILodVPJhNuaT 8zQg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543921; x=1756148721; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=XPu7HgEqjv/9pYy7ZOBzTh5eusrrnsbsNUri9J4cP3I=; b=QFggEDZPjv7ftgMcdIpaq9ycf6NhSrEzTwierq968fvyMSMX8mwMJhApp63pcFZ6Wz QPSVKk7h05oKkFPttUnte3h7hXEuhQRIPNefTRHqCyHsKpOR1siz9KWstkC1+N9Wx94m l2ctR1X8Osnnr+t3AnH9jRj86LK0fFumyxgDFKFrIoXwOEf62ufC64gPUkOhFNMEwhiP 4J2lHhDevmFluWL2fM38WLbfoFHymNCJobJtPV/o0TQkGnKCZ3oIMzkfRRCt3Jfku+RY lTeXak9fnKAIhIIwVl6AqC+74qNt7m6ke+en+hsX72RGvMaHudThzC8VL3XUqAepmQOG 0vnw== X-Forwarded-Encrypted: i=1; AJvYcCVFjpj188ZmvDhXRIaDUXeelZ0ZD9aQINqaOHmUZdbdcVj2L/bf9DPM8MMKAWOyxd+r0Ib/qnSEQB6azTQ=@vger.kernel.org X-Gm-Message-State: AOJu0YyoCe989Iv7mOmkOywgMIJwurWRbUjzOF7zjcGqBGCpKkwio/sH NcfgsE6TuFOXs7uNBWamYeGpxBRr8Bm7l4O1hRWa/D4mDd+mhBy//hQHcsjamZHkC38F3nF3i/r +DAgwr42scw== X-Google-Smtp-Source: AGHT+IGCajH84fWcoM3zhwef2bMwrEFrQjDsj6evWediiGXYFqCovJ4ispp6xFqBqm/wBfDbU9d2jJ2h/tqS X-Received: from pldq4.prod.google.com ([2002:a17:902:c9c4:b0:240:608a:938d]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:903:189:b0:244:214f:13b7 with SMTP id d9443c01a7336-24478feb7cdmr134225075ad.53.1755543921168; Mon, 18 Aug 2025 12:05:21 -0700 (PDT) Date: Mon, 18 Aug 2025 12:03:58 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-3-irogers@google.com> Subject: [PATCH v2 02/20] perf vendor events: Update alderlake events/metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update events from v1.31 to v1.33. Update metrics from TMA 5.0 to 5.1. The event updates come from: https://github.com/intel/perfmon/commit/c504da6cb00e52067206166d2049a0c11af= 3b650 https://github.com/intel/perfmon/commit/4c18312c1a22eb564f5dbc94b187b59e438= 3a56a Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/alderlake/adl-metrics.json | 104 +++++------ .../pmu-events/arch/x86/alderlake/cache.json | 99 +++++------ .../arch/x86/alderlake/floating-point.json | 28 ++- .../arch/x86/alderlake/frontend.json | 42 +++-- .../pmu-events/arch/x86/alderlake/memory.json | 12 +- .../pmu-events/arch/x86/alderlake/other.json | 8 +- .../arch/x86/alderlake/pipeline.json | 163 +++++++----------- .../x86/alderlake/uncore-interconnect.json | 2 - .../arch/x86/alderlake/virtual-memory.json | 40 ++--- .../arch/x86/alderlaken/adln-metrics.json | 20 +-- .../x86/alderlaken/uncore-interconnect.json | 2 - tools/perf/pmu-events/arch/x86/mapfile.csv | 4 +- 12 files changed, 231 insertions(+), 293 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/alderlake/adl-metrics.json b/to= ols/perf/pmu-events/arch/x86/alderlake/adl-metrics.json index 377dfecd96bd..cae7c0cf02f2 100644 --- a/tools/perf/pmu-events/arch/x86/alderlake/adl-metrics.json +++ b/tools/perf/pmu-events/arch/x86/alderlake/adl-metrics.json @@ -1,56 +1,56 @@ [ { "BriefDescription": "C10 residency percent per package", - "MetricExpr": "cstate_pkg@c10\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c10\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C10_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C8 residency percent per package", - "MetricExpr": "cstate_pkg@c8\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c8\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C8_Pkg_Residency", "ScaleUnit": "100%" @@ -552,7 +552,7 @@ }, { "BriefDescription": "Average CPU Utilization", - "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.REF_TSC@ / TSC", + "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.REF_TSC@ / msr@tsc\\,cpu= =3Dcpu_atom@", "MetricName": "tma_info_system_cpu_utilization", "Unit": "cpu_atom" }, @@ -751,7 +751,7 @@ { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-be\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-be\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BvOB;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", @@ -789,12 +789,21 @@ "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)", "Unit": "cpu_core" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: ", + "Unit": "cpu_core" + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)= ))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full", "Unit": "cpu_core" }, @@ -802,23 +811,14 @@ "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependen= cy + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_= bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma= _l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_dtlb_load + tma_fb= _full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tm= a_store_fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tm= a_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_split_l= oads / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_= latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_s= tore_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_sharin= g + tma_split_stores + tma_store_latency + tma_streaming_stores)) + tma_mem= ory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_boun= d + tma_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store= + tma_false_sharing + tma_split_stores + tma_store_latency + tma_streaming= _stores)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency", "Unit": "cpu_core" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: ", - "Unit": "cpu_core" - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - c= pu_core@INST_RETIRED.REP_ITERATION@ / cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D= 1@) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cl= ears_resteers + tma_mispredicts_resteers * tma_other_mispredicts / tma_bran= ch_mispredicts) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unk= nown_branches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_miss= es + tma_itlb_misses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * t= ma_ms / (tma_dsb + tma_lsd + tma_mite + tma_ms))) - tma_bottleneck_big_code= ", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - c= pu_core@INST_RETIRED.REP_ITERATION@ / cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D= 1@) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cl= ears_resteers + tma_mispredicts_resteers * tma_other_mispredicts / tma_bran= ch_mispredicts) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unk= nown_branches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_miss= es + tma_itlb_misses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottlene= ck_big_code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20", @@ -826,7 +826,7 @@ }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * ((1 - cpu_core@INST_RETIRED.REP_ITERATION@ / = cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_restee= rs * tma_other_mispredicts / tma_branch_mispredicts) / (tma_clears_resteers= + tma_mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers= + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_m= s_switches) + tma_fetch_bandwidth * tma_ms / (tma_dsb + tma_lsd + tma_mite = + tma_ms)) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_bra= nch_mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_n= ukes / tma_other_nukes + tma_core_bound * (tma_serializing_operation + cpu_= core@RS.EMPTY_RESOURCE@ / tma_info_thread_clks * tma_ports_utilized_0) / (t= ma_divider + tma_ports_utilization + tma_serializing_operation) + tma_micro= code_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) * (t= ma_assists / tma_microcode_sequencer) * tma_heavy_operations)", + "MetricExpr": "100 * ((1 - cpu_core@INST_RETIRED.REP_ITERATION@ / = cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_restee= rs * tma_other_mispredicts / tma_branch_mispredicts) / (tma_clears_resteers= + tma_mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers= + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_m= s_switches) + tma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredict= s / tma_branch_mispredicts * tma_branch_mispredicts + tma_machine_clears * = tma_other_nukes / tma_other_nukes + tma_core_bound * (tma_serializing_opera= tion + cpu_core@RS.EMPTY_RESOURCE@ / tma_info_thread_clks * tma_ports_utili= zed_0) / (tma_divider + tma_ports_utilization + tma_serializing_operation) = + tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequ= encer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -862,7 +862,7 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -879,7 +879,7 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU has wasted due to Branch Misprediction", - "MetricExpr": "cpu_core@topdown\\-br\\-mispredict@ / (cpu_core@top= down\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-re= tiring@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-br\\-mispredict@ / (cpu_core@top= down\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-re= tiring@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BadSpec;BrMispredicts;BvMP;TmaL2;TopdownL2;tma_L2_= group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_specula= tion > 0.15", @@ -992,7 +992,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to contested acces= ses", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(25 * tma_info_system_core_frequency * (cpu_core@ME= M_LOAD_L3_HIT_RETIRED.XSNP_FWD@ * (cpu_core@OCR.DEMAND_DATA_RD.L3_HIT.SNOOP= _HITM@ / (cpu_core@OCR.DEMAND_DATA_RD.L3_HIT.SNOOP_HITM@ + cpu_core@OCR.DEM= AND_DATA_RD.L3_HIT.SNOOP_HIT_WITH_FWD@))) + 24 * tma_info_system_core_frequ= ency * cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS@) * (1 + cpu_core@MEM_LOA= D_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2) / tma_info_thre= ad_clks", "MetricGroup": "BvMS;DataSharing;LockCont;Offcore;Snoop;TopdownL4;= tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", @@ -1109,7 +1108,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -1238,7 +1237,7 @@ }, { "BriefDescription": "This metric represents fraction of slots wher= e the CPU was retiring heavy-weight operations -- instructions that require= two or more uops or micro-coded sequences", - "MetricExpr": "cpu_core@topdown\\-heavy\\-ops@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-heavy\\-ops@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Retire;TmaL2;TopdownL2;tma_L2_group;tma_retiring_g= roup", "MetricName": "tma_heavy_operations", "MetricThreshold": "tma_heavy_operations > 0.1", @@ -1851,7 +1850,7 @@ "Unit": "cpu_core" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "cpu_core@UOPS_EXECUTED.THREAD@ / (cpu_core@UOPS_EXE= CUTED.CORE_CYCLES_GE_1@ / 2 if #SMT_on else cpu_core@UOPS_EXECUTED.THREAD\\= ,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute", @@ -1912,7 +1911,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc\\,cpu= =3Dcpu_core@ / 1e9 / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency", "Unit": "cpu_core" @@ -1926,7 +1925,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.REF_TSC@ / TSC", + "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.REF_TSC@ / msr@tsc\\,cpu= =3Dcpu_core@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized", "Unit": "cpu_core" @@ -1936,7 +1935,7 @@ "MetricExpr": "64 * (UNC_ARB_TRK_REQUESTS.ALL + UNC_ARB_COH_TRK_RE= QUESTS.ALL) / 1e6 / tma_info_system_time / 1e3", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full", + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full", "Unit": "cpu_core" }, { @@ -1980,7 +1979,6 @@ }, { "BriefDescription": "Average latency of data read request to exter= nal memory (in nanoseconds)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(UNC_ARB_TRK_OCCUPANCY.RD + UNC_ARB_DAT_OCCUPANCY.R= D) / UNC_ARB_TRK_REQUESTS.RD", "MetricGroup": "Mem;MemoryLat;SoC", "MetricName": "tma_info_system_mem_read_latency", @@ -2031,6 +2029,13 @@ "MetricName": "tma_info_system_turbo_utilization", "Unit": "cpu_core" }, + { + "BriefDescription": "Measured Average Uncore Frequency for the SoC= [GHz]", + "MetricExpr": "tma_info_system_socket_clks / 1e9 / tma_info_system= _time", + "MetricGroup": "SoC", + "MetricName": "tma_info_system_uncore_frequency", + "Unit": "cpu_core" + }, { "BriefDescription": "Per-Logical Processor actual clocks when the = Logical Processor is active.", "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.THREAD@", @@ -2150,12 +2155,12 @@ "Unit": "cpu_core" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", "MetricExpr": "min(2 * (cpu_core@MEM_INST_RETIRED.ALL_LOADS@ - cpu= _core@MEM_LOAD_RETIRED.FB_HIT@ - cpu_core@MEM_LOAD_RETIRED.L1_MISS@) * 20 /= 100, max(cpu_core@CYCLE_ACTIVITY.CYCLES_MEM_ANY@ - cpu_core@MEMORY_ACTIVIT= Y.CYCLES_L1D_MISS@, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2171,7 +2176,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles wit= h demand load accesses that hit the L2 cache under unloaded scenarios (poss= ibly L2 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "3 * tma_info_system_core_frequency * cpu_core@MEM_L= OAD_RETIRED.L2_HIT@ * (1 + cpu_core@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM= _LOAD_RETIRED.L1_MISS@ / 2) / tma_info_thread_clks", "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_l2_bound_grou= p", "MetricName": "tma_l2_hit_latency", @@ -2192,12 +2196,11 @@ }, { "BriefDescription": "This metric estimates fraction of cycles with= demand load accesses that hit the L3 cache under unloaded scenarios (possi= bly L3 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "9 * tma_info_system_core_frequency * (cpu_core@MEM_= LOAD_RETIRED.L3_HIT@ * (1 + cpu_core@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@ME= M_LOAD_RETIRED.L1_MISS@ / 2)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2279,6 +2282,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU spent handling cache misses due to lock operations", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(16 * max(0, cpu_core@MEM_INST_RETIRED.LOCK_LOADS@ = - cpu_core@L2_RQSTS.ALL_RFO@) + cpu_core@MEM_INST_RETIRED.LOCK_LOADS@ / cpu= _core@MEM_INST_RETIRED.ALL_STORES@ * (10 * cpu_core@L2_RQSTS.RFO_HIT@ + min= (cpu_core@CPU_CLK_UNHALTED.THREAD@, cpu_core@OFFCORE_REQUESTS_OUTSTANDING.C= YCLES_WITH_DEMAND_RFO@))) / tma_info_thread_clks", "MetricGroup": "LockCont;Offcore;TopdownL4;tma_L4_group;tma_issueR= FO;tma_l1_bound_group", "MetricName": "tma_lock_latency", @@ -2314,7 +2318,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2324,13 +2328,13 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%", "Unit": "cpu_core" }, { "BriefDescription": "This metric represents fraction of slots the = Memory subsystem within the Backend was a bottleneck", - "MetricExpr": "cpu_core@topdown\\-mem\\-bound@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-mem\\-bound@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Backend;TmaL2;TopdownL2;tma_L2_group;tma_backend_b= ound_group", "MetricName": "tma_memory_bound", "MetricThreshold": "tma_memory_bound > 0.2 & tma_backend_bound > 0= .2", @@ -2341,7 +2345,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to LFENCE Instructions.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * cpu_core@MISC2_RETIRED.LFENCE@ / tma_info_thre= ad_clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_memory_fence", @@ -2400,7 +2403,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "max(cpu_core@IDQ.MS_CYCLES_ANY@, cpu_core@UOPS_RETI= RED.MS\\,cmask\\=3D1@ / (cpu_core@UOPS_RETIRED.SLOTS@ / cpu_core@UOPS_ISSUE= D.ANY@)) / tma_info_core_core_clks / 2", + "MetricExpr": "max(cpu_core@IDQ.MS_CYCLES_ANY@, cpu_core@UOPS_RETI= RED.MS\\,cmask\\=3D1@ / (cpu_core@UOPS_RETIRED.SLOTS@ / cpu_core@UOPS_ISSUE= D.ANY@)) / tma_info_core_core_clks / 2.4", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -2439,6 +2442,7 @@ }, { "BriefDescription": "This metric represents the remaining light uo= ps fraction the CPU has executed - remaining means not covered by other sib= ling nodes", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_light_operations - (tma_fp_arith + tma_i= nt_operations + tma_memory_operations + tma_fused_instructions + tma_non_fu= sed_branches))", "MetricGroup": "Pipeline;TopdownL3;tma_L3_group;tma_light_operatio= ns_group", "MetricName": "tma_other_light_ops", @@ -2507,6 +2511,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "((tma_ports_utilized_0 * tma_info_thread_clks + (cp= u_core@EXE_ACTIVITY.1_PORTS_UTIL@ + tma_retiring * cpu_core@EXE_ACTIVITY.2_= 3_PORTS_UTIL@)) / tma_info_thread_clks if cpu_core@ARITH.DIV_ACTIVE@ < cpu_= core@CYCLE_ACTIVITY.STALLS_TOTAL@ - cpu_core@EXE_ACTIVITY.BOUND_ON_LOADS@ e= lse (cpu_core@EXE_ACTIVITY.1_PORTS_UTIL@ + tma_retiring * cpu_core@EXE_ACTI= VITY.2_3_PORTS_UTIL@) / tma_info_thread_clks)", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", @@ -2517,6 +2522,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "(cpu_core@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ + max(cpu= _core@RS.EMPTY_RESOURCE@ - cpu_core@RESOURCE_STALLS.SCOREBOARD@, 0)) / tma_= info_thread_clks * (cpu_core@CYCLE_ACTIVITY.STALLS_TOTAL@ - cpu_core@EXE_AC= TIVITY.BOUND_ON_LOADS@) / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", @@ -2527,6 +2533,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "cpu_core@EXE_ACTIVITY.1_PORTS_UTIL@ / tma_info_thre= ad_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", @@ -2537,7 +2544,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@EXE_ACTIVITY.2_PORTS_UTIL@ / tma_info_thre= ad_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", @@ -2548,7 +2554,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 3 or more uops per cycle on all execution ports (Logical= Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@UOPS_EXECUTED.CYCLES_GE_3@ / tma_info_thre= ad_clks", "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_ut= ilization_group", "MetricName": "tma_ports_utilized_3m", @@ -2560,7 +2565,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-retiring@ / (cpu_core@topdown\\-= fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@= + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-retiring@ / (cpu_core@topdown\\-= fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@= + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -2591,7 +2596,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to PAUSE Instructions", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.PAUSE@ / tma_info_thread_= clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_slow_pause", @@ -2626,7 +2630,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%", "Unit": "cpu_core" }, diff --git a/tools/perf/pmu-events/arch/x86/alderlake/cache.json b/tools/pe= rf/pmu-events/arch/x86/alderlake/cache.json index 5461576dafc7..6a56c9ad8e43 100644 --- a/tools/perf/pmu-events/arch/x86/alderlake/cache.json +++ b/tools/perf/pmu-events/arch/x86/alderlake/cache.json @@ -4,7 +4,6 @@ "Counter": "0,1,2,3", "EventCode": "0x51", "EventName": "L1D.HWPF_MISS", - "PublicDescription": "L1D.HWPF_MISS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x20", "Unit": "cpu_core" @@ -14,7 +13,7 @@ "Counter": "0,1,2,3", "EventCode": "0x51", "EventName": "L1D.REPLACEMENT", - "PublicDescription": "Counts L1D data line replacements including = opportunistic replacements, and replacements that require stall-for-replace= or block-for-replace. Available PDIST counters: 0", + "PublicDescription": "Counts L1D data line replacements including = opportunistic replacements, and replacements that require stall-for-replace= or block-for-replace.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -24,7 +23,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.FB_FULL", - "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -36,7 +35,7 @@ "EdgeDetect": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.FB_FULL_PERIODS", - "PublicDescription": "Counts number of phases a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts number of phases a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -47,7 +46,6 @@ "Deprecated": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.L2_STALL", - "PublicDescription": "This event is deprecated. Refer to new event= L1D_PEND_MISS.L2_STALLS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x4", "Unit": "cpu_core" @@ -57,7 +55,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.L2_STALLS", - "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D due to lack of L2 resources. Demand requests include cac= heable/uncacheable demand load, store, lock or SW prefetch accesses. Availa= ble PDIST counters: 0", + "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D due to lack of L2 resources. Demand requests include cac= heable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x4", "Unit": "cpu_core" @@ -67,7 +65,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.PENDING", - "PublicDescription": "Counts number of L1D misses that are outstan= ding in each cycle, that is each cycle the number of Fill Buffers (FB) outs= tanding required by Demand Reads. FB either is held by demand loads, or it = is held by non-demand loads and gets hit at least once by demand. The valid= outstanding interval is defined until the FB deallocation by one of the fo= llowing ways: from FB allocation, if FB is allocated by demand from the dem= and Hit FB, if it is allocated by hardware or software prefetch. Note: In t= he L1D, a Demand Read contains cacheable or noncacheable demand loads, incl= uding ones causing cache-line splits and reads due to page walks resulted f= rom any request type. Available PDIST counters: 0", + "PublicDescription": "Counts number of L1D misses that are outstan= ding in each cycle, that is each cycle the number of Fill Buffers (FB) outs= tanding required by Demand Reads. FB either is held by demand loads, or it = is held by non-demand loads and gets hit at least once by demand. The valid= outstanding interval is defined until the FB deallocation by one of the fo= llowing ways: from FB allocation, if FB is allocated by demand from the dem= and Hit FB, if it is allocated by hardware or software prefetch. Note: In t= he L1D, a Demand Read contains cacheable or noncacheable demand loads, incl= uding ones causing cache-line splits and reads due to page walks resulted f= rom any request type.", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -78,7 +76,7 @@ "CounterMask": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.PENDING_CYCLES", - "PublicDescription": "Counts duration of L1D miss outstanding in c= ycles. Available PDIST counters: 0", + "PublicDescription": "Counts duration of L1D miss outstanding in c= ycles.", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -88,7 +86,7 @@ "Counter": "0,1,2,3", "EventCode": "0x25", "EventName": "L2_LINES_IN.ALL", - "PublicDescription": "Counts the number of L2 cache lines filling = the L2. Counting does not cover rejects. Available PDIST counters: 0", + "PublicDescription": "Counts the number of L2 cache lines filling = the L2. Counting does not cover rejects.", "SampleAfterValue": "100003", "UMask": "0x1f", "Unit": "cpu_core" @@ -98,7 +96,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.NON_SILENT", - "PublicDescription": "Counts the number of lines that are evicted = by L2 cache when triggered by an L2 cache fill. Those lines are in Modified= state. Modified lines are written back to L3 Available PDIST counters: 0", + "PublicDescription": "Counts the number of lines that are evicted = by L2 cache when triggered by an L2 cache fill. Those lines are in Modified= state. Modified lines are written back to L3", "SampleAfterValue": "200003", "UMask": "0x2", "Unit": "cpu_core" @@ -108,7 +106,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.SILENT", - "PublicDescription": "Counts the number of lines that are silently= dropped by L2 cache. These lines are typically in Shared or Exclusive stat= e. A non-threaded event. Available PDIST counters: 0", + "PublicDescription": "Counts the number of lines that are silently= dropped by L2 cache. These lines are typically in Shared or Exclusive stat= e. A non-threaded event.", "SampleAfterValue": "200003", "UMask": "0x1", "Unit": "cpu_core" @@ -118,7 +116,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.USELESS_HWPF", - "PublicDescription": "Counts the number of cache lines that have b= een prefetched by the L2 hardware prefetcher but not used by demand access = when evicted from the L2 cache Available PDIST counters: 0", + "PublicDescription": "Counts the number of cache lines that have b= een prefetched by the L2 hardware prefetcher but not used by demand access = when evicted from the L2 cache", "SampleAfterValue": "200003", "UMask": "0x4", "Unit": "cpu_core" @@ -137,7 +135,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.ALL", - "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_RQSTS.REFERENCES] Available PDIST coun= ters: 0", + "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_RQSTS.REFERENCES]", "SampleAfterValue": "200003", "UMask": "0xff", "Unit": "cpu_core" @@ -167,7 +165,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.MISS", - "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_RQSTS.MISS] Available PDIST coun= ters: 0", + "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_RQSTS.MISS]", "SampleAfterValue": "200003", "UMask": "0x3f", "Unit": "cpu_core" @@ -177,7 +175,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_CODE_RD", - "PublicDescription": "Counts the total number of L2 code requests.= Available PDIST counters: 0", + "PublicDescription": "Counts the total number of L2 code requests.= ", "SampleAfterValue": "200003", "UMask": "0xe4", "Unit": "cpu_core" @@ -187,7 +185,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_DATA_RD", - "PublicDescription": "Counts Demand Data Read requests accessing t= he L2 cache. These requests may hit or miss L2 cache. True-miss exclude mis= ses that were merged with ongoing L2 misses. An access is counted once. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts Demand Data Read requests accessing t= he L2 cache. These requests may hit or miss L2 cache. True-miss exclude mis= ses that were merged with ongoing L2 misses. An access is counted once.", "SampleAfterValue": "200003", "UMask": "0xe1", "Unit": "cpu_core" @@ -197,7 +195,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_MISS", - "PublicDescription": "Counts demand requests that miss L2 cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts demand requests that miss L2 cache.", "SampleAfterValue": "200003", "UMask": "0x27", "Unit": "cpu_core" @@ -207,7 +205,6 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_HWPF", - "PublicDescription": "L2_RQSTS.ALL_HWPF Available PDIST counters: = 0", "SampleAfterValue": "200003", "UMask": "0xf0", "Unit": "cpu_core" @@ -217,7 +214,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_RFO", - "PublicDescription": "Counts the total number of RFO (read for own= ership) requests to L2 cache. L2 RFO requests include both L1D demand RFO m= isses as well as L1D RFO prefetches. Available PDIST counters: 0", + "PublicDescription": "Counts the total number of RFO (read for own= ership) requests to L2 cache. L2 RFO requests include both L1D demand RFO m= isses as well as L1D RFO prefetches.", "SampleAfterValue": "200003", "UMask": "0xe2", "Unit": "cpu_core" @@ -227,7 +224,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.CODE_RD_HIT", - "PublicDescription": "Counts L2 cache hits when fetching instructi= ons, code reads. Available PDIST counters: 0", + "PublicDescription": "Counts L2 cache hits when fetching instructi= ons, code reads.", "SampleAfterValue": "200003", "UMask": "0xc4", "Unit": "cpu_core" @@ -237,7 +234,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.CODE_RD_MISS", - "PublicDescription": "Counts L2 cache misses when fetching instruc= tions. Available PDIST counters: 0", + "PublicDescription": "Counts L2 cache misses when fetching instruc= tions.", "SampleAfterValue": "200003", "UMask": "0x24", "Unit": "cpu_core" @@ -247,7 +244,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT", - "PublicDescription": "Counts the number of demand Data Read reques= ts initiated by load instructions that hit L2 cache. Available PDIST counte= rs: 0", + "PublicDescription": "Counts the number of demand Data Read reques= ts initiated by load instructions that hit L2 cache.", "SampleAfterValue": "200003", "UMask": "0xc1", "Unit": "cpu_core" @@ -257,7 +254,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.DEMAND_DATA_RD_MISS", - "PublicDescription": "Counts demand Data Read requests with true-m= iss in the L2 cache. True-miss excludes misses that were merged with ongoin= g L2 misses. An access is counted once. Available PDIST counters: 0", + "PublicDescription": "Counts demand Data Read requests with true-m= iss in the L2 cache. True-miss excludes misses that were merged with ongoin= g L2 misses. An access is counted once.", "SampleAfterValue": "200003", "UMask": "0x21", "Unit": "cpu_core" @@ -267,7 +264,6 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.HWPF_MISS", - "PublicDescription": "L2_RQSTS.HWPF_MISS Available PDIST counters:= 0", "SampleAfterValue": "200003", "UMask": "0x30", "Unit": "cpu_core" @@ -277,7 +273,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.MISS", - "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_REQUEST.MISS] Available PDIST co= unters: 0", + "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_REQUEST.MISS]", "SampleAfterValue": "200003", "UMask": "0x3f", "Unit": "cpu_core" @@ -287,7 +283,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.REFERENCES", - "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_REQUEST.ALL] Available PDIST counters:= 0", + "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_REQUEST.ALL]", "SampleAfterValue": "200003", "UMask": "0xff", "Unit": "cpu_core" @@ -297,7 +293,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.RFO_HIT", - "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that hit L2 cache. Available PDIST counters: 0", + "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that hit L2 cache.", "SampleAfterValue": "200003", "UMask": "0xc2", "Unit": "cpu_core" @@ -307,7 +303,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.RFO_MISS", - "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that miss L2 cache. Available PDIST counters: 0", + "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that miss L2 cache.", "SampleAfterValue": "200003", "UMask": "0x22", "Unit": "cpu_core" @@ -317,7 +313,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.SWPF_HIT", - "PublicDescription": "Counts Software prefetch requests that hit t= he L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when = FB is not full. Available PDIST counters: 0", + "PublicDescription": "Counts Software prefetch requests that hit t= he L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when = FB is not full.", "SampleAfterValue": "200003", "UMask": "0xc8", "Unit": "cpu_core" @@ -327,7 +323,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.SWPF_MISS", - "PublicDescription": "Counts Software prefetch requests that miss = the L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when= FB is not full. Available PDIST counters: 0", + "PublicDescription": "Counts Software prefetch requests that miss = the L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when= FB is not full.", "SampleAfterValue": "200003", "UMask": "0x28", "Unit": "cpu_core" @@ -337,7 +333,7 @@ "Counter": "0,1,2,3", "EventCode": "0x23", "EventName": "L2_TRANS.L2_WB", - "PublicDescription": "Counts L2 writebacks that access L2 cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts L2 writebacks that access L2 cache.", "SampleAfterValue": "200003", "UMask": "0x40", "Unit": "cpu_core" @@ -357,7 +353,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x2e", "EventName": "LONGEST_LAT_CACHE.MISS", - "PublicDescription": "Counts core-originated cacheable requests th= at miss the L3 cache (Longest Latency cache). Requests include data and cod= e reads, Reads-for-Ownership (RFOs), speculative accesses and hardware pref= etches to the L1 and L2. It does not include hardware prefetches to the L3= , and may not count other types of requests to the L3. Available PDIST coun= ters: 0", + "PublicDescription": "Counts core-originated cacheable requests th= at miss the L3 cache (Longest Latency cache). Requests include data and cod= e reads, Reads-for-Ownership (RFOs), speculative accesses and hardware pref= etches to the L1 and L2. It does not include hardware prefetches to the L3= , and may not count other types of requests to the L3.", "SampleAfterValue": "100003", "UMask": "0x41", "Unit": "cpu_core" @@ -377,7 +373,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x2e", "EventName": "LONGEST_LAT_CACHE.REFERENCE", - "PublicDescription": "Counts core-originated cacheable requests to= the L3 cache (Longest Latency cache). Requests include data and code reads= , Reads-for-Ownership (RFOs), speculative accesses and hardware prefetches = to the L1 and L2. It does not include hardware prefetches to the L3, and m= ay not count other types of requests to the L3. Available PDIST counters: 0= ", + "PublicDescription": "Counts core-originated cacheable requests to= the L3 cache (Longest Latency cache). Requests include data and code reads= , Reads-for-Ownership (RFOs), speculative accesses and hardware prefetches = to the L1 and L2. It does not include hardware prefetches to the L3, and m= ay not count other types of requests to the L3.", "SampleAfterValue": "100003", "UMask": "0x4f", "Unit": "cpu_core" @@ -552,7 +548,7 @@ "Counter": "0,1,2,3", "EventCode": "0x43", "EventName": "MEM_LOAD_COMPLETED.L1_MISS_ANY", - "PublicDescription": "Number of completed demand load requests tha= t missed the L1 data cache including shadow misses (FB hits, merge to an on= going L1D miss) Available PDIST counters: 0", + "PublicDescription": "Number of completed demand load requests tha= t missed the L1 data cache including shadow misses (FB hits, merge to an on= going L1D miss)", "SampleAfterValue": "1000003", "UMask": "0xfd", "Unit": "cpu_core" @@ -853,7 +849,6 @@ "Counter": "0,1,2,3", "EventCode": "0x44", "EventName": "MEM_STORE_RETIRED.L2_HIT", - "PublicDescription": "MEM_STORE_RETIRED.L2_HIT Available PDIST cou= nters: 0", "SampleAfterValue": "200003", "UMask": "0x1", "Unit": "cpu_core" @@ -1050,7 +1045,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe5", "EventName": "MEM_UOP_RETIRED.ANY", - "PublicDescription": "Number of retired micro-operations (uops) fo= r load or store memory accesses Available PDIST counters: 0", + "PublicDescription": "Number of retired micro-operations (uops) fo= r load or store memory accesses", "SampleAfterValue": "1000003", "UMask": "0x3", "Unit": "cpu_core" @@ -1372,7 +1367,6 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.ALL_REQUESTS", - "PublicDescription": "OFFCORE_REQUESTS.ALL_REQUESTS Available PDIS= T counters: 0", "SampleAfterValue": "100003", "UMask": "0x80", "Unit": "cpu_core" @@ -1382,7 +1376,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DATA_RD", - "PublicDescription": "Counts the demand and prefetch data reads. A= ll Core Data Reads include cacheable 'Demands' and L2 prefetchers (not L3 p= refetchers). Counting also covers reads due to page walks resulted from any= request type. Available PDIST counters: 0", + "PublicDescription": "Counts the demand and prefetch data reads. A= ll Core Data Reads include cacheable 'Demands' and L2 prefetchers (not L3 p= refetchers). Counting also covers reads due to page walks resulted from any= request type.", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -1392,7 +1386,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_CODE_RD", - "PublicDescription": "Counts both cacheable and non-cacheable code= read requests. Available PDIST counters: 0", + "PublicDescription": "Counts both cacheable and non-cacheable code= read requests.", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -1402,7 +1396,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_DATA_RD", - "PublicDescription": "Counts the Demand Data Read requests sent to= uncore. Use it in conjunction with OFFCORE_REQUESTS_OUTSTANDING to determi= ne average latency in the uncore. Available PDIST counters: 0", + "PublicDescription": "Counts the Demand Data Read requests sent to= uncore. Use it in conjunction with OFFCORE_REQUESTS_OUTSTANDING to determi= ne average latency in the uncore.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -1412,7 +1406,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_RFO", - "PublicDescription": "Counts the demand RFO (read for ownership) r= equests including regular RFOs, locks, ItoM. Available PDIST counters: 0", + "PublicDescription": "Counts the demand RFO (read for ownership) r= equests including regular RFOs, locks, ItoM.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -1424,7 +1418,6 @@ "Errata": "ADL038", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD", - "PublicDescription": "This event is deprecated. Refer to new event= OFFCORE_REQUESTS_OUTSTANDING.DATA_RD Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -1436,7 +1429,6 @@ "Errata": "ADL038", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD", - "PublicDescription": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DAT= A_RD Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -1447,7 +1439,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_CODE= _RD", - "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1458,7 +1450,6 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA= _RD", - "PublicDescription": "Cycles where at least 1 outstanding demand d= ata read request is pending. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1469,7 +1460,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO", - "PublicDescription": "Counts the number of offcore outstanding dem= and rfo Reads transactions in the super queue every cycle. The 'Offcore out= standing' state of the transaction lasts from the L2 miss until the sending= transaction completion to requestor (SQ deallocation). See the correspondi= ng Umask under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding dem= and rfo Reads transactions in the super queue every cycle. The 'Offcore out= standing' state of the transaction lasts from the L2 miss until the sending= transaction completion to requestor (SQ deallocation). See the correspondi= ng Umask under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x4", "Unit": "cpu_core" @@ -1480,7 +1471,6 @@ "Errata": "ADL038", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DATA_RD", - "PublicDescription": "OFFCORE_REQUESTS_OUTSTANDING.DATA_RD Availab= le PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -1490,7 +1480,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_CODE_RD", - "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1500,7 +1490,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD", - "PublicDescription": "For every cycle, increments by the number of= outstanding demand data read requests pending. Requests are considered o= utstanding from the time they miss the core's L2 cache until the transactio= n completion message is sent to the requestor. Available PDIST counters: 0", + "PublicDescription": "For every cycle, increments by the number of= outstanding demand data read requests pending. Requests are considered o= utstanding from the time they miss the core's L2 cache until the transactio= n completion message is sent to the requestor.", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1510,7 +1500,7 @@ "Counter": "0,1,2,3", "EventCode": "0x2c", "EventName": "SQ_MISC.BUS_LOCK", - "PublicDescription": "Counts the more expensive bus lock needed to= enforce cache coherency for certain memory accesses that need to be done a= tomically. Can be created by issuing an atomic instruction (via the LOCK p= refix) which causes a cache line split or accesses uncacheable memory. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts the more expensive bus lock needed to= enforce cache coherency for certain memory accesses that need to be done a= tomically. Can be created by issuing an atomic instruction (via the LOCK p= refix) which causes a cache line split or accesses uncacheable memory.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -1520,7 +1510,6 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.ANY", - "PublicDescription": "Counts the number of PREFETCHNTA, PREFETCHW,= PREFETCHT0, PREFETCHT1 or PREFETCHT2 instructions executed. Available PDIS= T counters: 0", "SampleAfterValue": "100003", "UMask": "0xf", "Unit": "cpu_core" @@ -1530,7 +1519,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.NTA", - "PublicDescription": "Counts the number of PREFETCHNTA instruction= s executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHNTA instruction= s executed.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -1540,7 +1529,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.PREFETCHW", - "PublicDescription": "Counts the number of PREFETCHW instructions = executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHW instructions = executed.", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -1550,7 +1539,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.T0", - "PublicDescription": "Counts the number of PREFETCHT0 instructions= executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHT0 instructions= executed.", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -1560,7 +1549,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.T1_T2", - "PublicDescription": "Counts the number of PREFETCHT1 or PREFETCHT= 2 instructions executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHT1 or PREFETCHT= 2 instructions executed.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/alderlake/floating-point.json b= /tools/perf/pmu-events/arch/x86/alderlake/floating-point.json index d01f1b163ed8..62fd70f220e5 100644 --- a/tools/perf/pmu-events/arch/x86/alderlake/floating-point.json +++ b/tools/perf/pmu-events/arch/x86/alderlake/floating-point.json @@ -14,7 +14,6 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.FPDIV_ACTIVE", - "PublicDescription": "ARITH.FPDIV_ACTIVE Available PDIST counters:= 0", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -33,7 +32,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.FP", - "PublicDescription": "Counts all microcode Floating Point assists.= Available PDIST counters: 0", + "PublicDescription": "Counts all microcode Floating Point assists.= ", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -43,7 +42,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.SSE_AVX_MIX", - "PublicDescription": "ASSISTS.SSE_AVX_MIX Available PDIST counters= : 0", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -53,7 +51,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_0", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_0 [This event is al= ias to FP_ARITH_DISPATCHED.V0] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -63,7 +60,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_1", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_1 [This event is al= ias to FP_ARITH_DISPATCHED.V1] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -73,7 +69,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_5", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_5 [This event is al= ias to FP_ARITH_DISPATCHED.V2] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -83,7 +78,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V0", - "PublicDescription": "FP_ARITH_DISPATCHED.V0 [This event is alias = to FP_ARITH_DISPATCHED.PORT_0] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -93,7 +87,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V1", - "PublicDescription": "FP_ARITH_DISPATCHED.V1 [This event is alias = to FP_ARITH_DISPATCHED.PORT_1] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -103,7 +96,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V2", - "PublicDescription": "FP_ARITH_DISPATCHED.V2 [This event is alias = to FP_ARITH_DISPATCHED.PORT_5] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -113,7 +105,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 2 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as the= y perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR re= gister need to be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 2 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as the= y perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR re= gister need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -123,7 +115,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events. Available PDIST co= unters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -133,7 +125,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 = calculations per element. The DAZ and FTZ flags in the MXCSR register need = to be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 = calculations per element. The DAZ and FTZ flags in the MXCSR register need = to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -143,7 +135,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events. Available PDIST co= unters: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -153,7 +145,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.4_FLOPS", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision and 256-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 2 or/and 4 computation operations, one for each element. = Applies to SSE* and AVX* packed single precision floating-point and packed= double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL= DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB ins= tructions count twice as they perform 2 calculations per element. The DAZ a= nd FTZ flags in the MXCSR register need to be set when using these events. = Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision and 256-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 2 or/and 4 computation operations, one for each element. = Applies to SSE* and AVX* packed single precision floating-point and packed= double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL= DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB ins= tructions count twice as they perform 2 calculations per element. The DAZ a= nd FTZ flags in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x18", "Unit": "cpu_core" @@ -163,7 +155,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR", - "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision and double precision floating-point instructions retired; some = instructions will count twice as noted below. Each count represents 1 comp= utational operation. Applies to SSE* and AVX* scalar single precision float= ing-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB= . FM(N)ADD/SUB instructions count twice as they perform 2 calculations per= element. The DAZ and FTZ flags in the MXCSR register need to be set when u= sing these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision and double precision floating-point instructions retired; some = instructions will count twice as noted below. Each count represents 1 comp= utational operation. Applies to SSE* and AVX* scalar single precision float= ing-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB= . FM(N)ADD/SUB instructions count twice as they perform 2 calculations per= element. The DAZ and FTZ flags in the MXCSR register need to be set when u= sing these events.", "SampleAfterValue": "1000003", "UMask": "0x3", "Unit": "cpu_core" @@ -173,7 +165,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational scalar doubl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar double precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions co= unt twice as they perform 2 calculations per element. The DAZ and FTZ flags= in the MXCSR register need to be set when using these events. Available PD= IST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar doubl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar double precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions co= unt twice as they perform 2 calculations per element. The DAZ and FTZ flags= in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -183,7 +175,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE", - "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar single precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instr= uctions count twice as they perform 2 calculations per element. The DAZ and= FTZ flags in the MXCSR register need to be set when using these events. Av= ailable PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar single precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instr= uctions count twice as they perform 2 calculations per element. The DAZ and= FTZ flags in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -193,7 +185,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.VECTOR", - "PublicDescription": "Number of any Vector retired FP arithmetic i= nstructions. The DAZ and FTZ flags in the MXCSR register need to be set wh= en using these events. Available PDIST counters: 0", + "PublicDescription": "Number of any Vector retired FP arithmetic i= nstructions. The DAZ and FTZ flags in the MXCSR register need to be set wh= en using these events.", "SampleAfterValue": "1000003", "UMask": "0xfc", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/alderlake/frontend.json b/tools= /perf/pmu-events/arch/x86/alderlake/frontend.json index dae3174a74fb..ff3b30c2619a 100644 --- a/tools/perf/pmu-events/arch/x86/alderlake/frontend.json +++ b/tools/perf/pmu-events/arch/x86/alderlake/frontend.json @@ -14,7 +14,7 @@ "Counter": "0,1,2,3", "EventCode": "0x60", "EventName": "BACLEARS.ANY", - "PublicDescription": "Number of times the front-end is resteered w= hen it finds a branch instruction in a fetch line. This is called Unknown B= ranch which occurs for the first time a branch instruction is fetched or wh= en the branch is not tracked by the BPU (Branch Prediction Unit) anymore. A= vailable PDIST counters: 0", + "PublicDescription": "Number of times the front-end is resteered w= hen it finds a branch instruction in a fetch line. This is called Unknown B= ranch which occurs for the first time a branch instruction is fetched or wh= en the branch is not tracked by the BPU (Branch Prediction Unit) anymore.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -24,7 +24,7 @@ "Counter": "0,1,2,3", "EventCode": "0x87", "EventName": "DECODE.LCP", - "PublicDescription": "Counts cycles that the Instruction Length de= coder (ILD) stalls occurred due to dynamically changing prefix length of th= e decoded instruction (by operand size prefix instruction 0x66, address siz= e prefix instruction 0x67 or REX.W for Intel64). Count is proportional to t= he number of prefixes in a 16B-line. This may result in a three-cycle penal= ty for each LCP (Length changing prefix) in a 16-byte chunk. Available PDIS= T counters: 0", + "PublicDescription": "Counts cycles that the Instruction Length de= coder (ILD) stalls occurred due to dynamically changing prefix length of th= e decoded instruction (by operand size prefix instruction 0x66, address siz= e prefix instruction 0x67 or REX.W for Intel64). Count is proportional to t= he number of prefixes in a 16B-line. This may result in a three-cycle penal= ty for each LCP (Length changing prefix) in a 16-byte chunk.", "SampleAfterValue": "500009", "UMask": "0x1", "Unit": "cpu_core" @@ -34,7 +34,6 @@ "Counter": "0,1,2,3", "EventCode": "0x87", "EventName": "DECODE.MS_BUSY", - "PublicDescription": "Cycles the Microcode Sequencer is busy. Avai= lable PDIST counters: 0", "SampleAfterValue": "500009", "UMask": "0x2", "Unit": "cpu_core" @@ -44,7 +43,7 @@ "Counter": "0,1,2,3", "EventCode": "0x61", "EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES", - "PublicDescription": "Decode Stream Buffer (DSB) is a Uop-cache th= at holds translations of previously fetched instructions that were decoded = by the legacy x86 decode pipeline (MITE). This event counts fetch penalty c= ycles when a transition occurs from DSB to MITE. Available PDIST counters: = 0", + "PublicDescription": "Decode Stream Buffer (DSB) is a Uop-cache th= at holds translations of previously fetched instructions that were decoded = by the legacy x86 decode pipeline (MITE). This event counts fetch penalty c= ycles when a transition occurs from DSB to MITE.", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -302,7 +301,7 @@ "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE_DATA.STALLS", - "PublicDescription": "Counts cycles where a code line fetch is sta= lled due to an L1 instruction cache miss. The decode pipeline works at a 32= Byte granularity. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where a code line fetch is sta= lled due to an L1 instruction cache miss. The decode pipeline works at a 32= Byte granularity.", "SampleAfterValue": "500009", "UMask": "0x4", "Unit": "cpu_core" @@ -314,7 +313,6 @@ "EdgeDetect": "1", "EventCode": "0x80", "EventName": "ICACHE_DATA.STALL_PERIODS", - "PublicDescription": "ICACHE_DATA.STALL_PERIODS Available PDIST co= unters: 0", "SampleAfterValue": "500009", "UMask": "0x4", "Unit": "cpu_core" @@ -324,7 +322,7 @@ "Counter": "0,1,2,3", "EventCode": "0x83", "EventName": "ICACHE_TAG.STALLS", - "PublicDescription": "Counts cycles where a code fetch is stalled = due to L1 instruction cache tag miss. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where a code fetch is stalled = due to L1 instruction cache tag miss.", "SampleAfterValue": "200003", "UMask": "0x4", "Unit": "cpu_core" @@ -335,7 +333,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES_ANY", - "PublicDescription": "Counts the number of cycles uops were delive= red to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) p= ath. Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles uops were delive= red to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) p= ath.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -346,7 +344,7 @@ "CounterMask": "6", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES_OK", - "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the D= SB (Decode Stream Buffer) path. Count includes uops that may 'bypass' the I= DQ. Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the D= SB (Decode Stream Buffer) path. Count includes uops that may 'bypass' the I= DQ.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -356,7 +354,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.DSB_UOPS", - "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Availab= le PDIST counters: 0", + "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -367,7 +365,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES_ANY", - "PublicDescription": "Counts the number of cycles uops were delive= red to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipe= line) path. During these cycles uops are not being delivered from the Decod= e Stream Buffer (DSB). Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles uops were delive= red to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipe= line) path. During these cycles uops are not being delivered from the Decod= e Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -378,7 +376,7 @@ "CounterMask": "6", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES_OK", - "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the M= ITE (legacy decode pipeline) path. During these cycles uops are not being d= elivered from the Decode Stream Buffer (DSB). Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the M= ITE (legacy decode pipeline) path. During these cycles uops are not being d= elivered from the Decode Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -388,7 +386,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_UOPS", - "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the MITE path. This also means that uops are= not being delivered from the Decode Stream Buffer (DSB). Available PDIST c= ounters: 0", + "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the MITE path. This also means that uops are= not being delivered from the Decode Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -399,7 +397,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_CYCLES_ANY", - "PublicDescription": "Counts cycles during which uops are being de= livered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS= ) is busy. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts cycles during which uops are being de= livered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS= ) is busy. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.", "SampleAfterValue": "2000003", "UMask": "0x20", "Unit": "cpu_core" @@ -411,7 +409,7 @@ "EdgeDetect": "1", "EventCode": "0x79", "EventName": "IDQ.MS_SWITCHES", - "PublicDescription": "Number of switches from DSB (Decode Stream B= uffer) or MITE (legacy decode pipeline) to the Microcode Sequencer. Availab= le PDIST counters: 0", + "PublicDescription": "Number of switches from DSB (Decode Stream B= uffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -421,7 +419,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_UOPS", - "PublicDescription": "Counts the total number of uops delivered by= the Microcode Sequencer (MS). Available PDIST counters: 0", + "PublicDescription": "Counts the total number of uops delivered by= the Microcode Sequencer (MS).", "SampleAfterValue": "1000003", "UMask": "0x20", "Unit": "cpu_core" @@ -431,7 +429,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CORE", - "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CORE] Available PDI= ST counters: 0", + "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -442,7 +440,7 @@ "CounterMask": "6", "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CYCLES_0_UOPS_DELIV.CORE", - "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CYCLES= _0_UOPS_DELIV.CORE] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CYCLES= _0_UOPS_DELIV.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -454,7 +452,7 @@ "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CYCLES_FE_WAS_OK", "Invert": "1", - "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_UOPS_N= OT_DELIVERED.CYCLES_FE_WAS_OK] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_UOPS_N= OT_DELIVERED.CYCLES_FE_WAS_OK]", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -464,7 +462,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CORE", - "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_BUBBLES.CORE] Available PDIST counters= : 0", + "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_BUBBLES.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -475,7 +473,7 @@ "CounterMask": "6", "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE", - "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_BUBBLES.CYCLES_0_UOPS_DEL= IV.CORE] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_BUBBLES.CYCLES_0_UOPS_DEL= IV.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -487,7 +485,7 @@ "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK", "Invert": "1", - "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_BUBBLE= S.CYCLES_FE_WAS_OK] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_BUBBLE= S.CYCLES_FE_WAS_OK]", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/alderlake/memory.json b/tools/p= erf/pmu-events/arch/x86/alderlake/memory.json index 07f5786bdbc0..a0260d5b8619 100644 --- a/tools/perf/pmu-events/arch/x86/alderlake/memory.json +++ b/tools/perf/pmu-events/arch/x86/alderlake/memory.json @@ -5,7 +5,6 @@ "CounterMask": "6", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L3_MISS", - "PublicDescription": "Execution stalls while L3 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x6", "Unit": "cpu_core" @@ -79,7 +78,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.MEMORY_ORDERING", - "PublicDescription": "Counts the number of Machine Clears detected= dye to memory ordering. Memory Ordering Machine Clears may apply when a me= mory read may not conform to the memory ordering rules of the x86 architect= ure Available PDIST counters: 0", + "PublicDescription": "Counts the number of Machine Clears detected= dye to memory ordering. Memory Ordering Machine Clears may apply when a me= mory read may not conform to the memory ordering rules of the x86 architect= ure", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -90,7 +89,6 @@ "CounterMask": "2", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.CYCLES_L1D_MISS", - "PublicDescription": "Cycles while L1 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -101,7 +99,6 @@ "CounterMask": "3", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L1D_MISS", - "PublicDescription": "Execution stalls while L1 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x3", "Unit": "cpu_core" @@ -112,7 +109,7 @@ "CounterMask": "5", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L2_MISS", - "PublicDescription": "Execution stalls while L2 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock). Available PDIST counters: 0", + "PublicDescription": "Execution stalls while L2 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock).", "SampleAfterValue": "1000003", "UMask": "0x5", "Unit": "cpu_core" @@ -123,7 +120,7 @@ "CounterMask": "9", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L3_MISS", - "PublicDescription": "Execution stalls while L3 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock). Available PDIST counters: 0", + "PublicDescription": "Execution stalls while L3 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock).", "SampleAfterValue": "1000003", "UMask": "0x9", "Unit": "cpu_core" @@ -417,7 +414,6 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.L3_MISS_DEMAND_DATA_RD", - "PublicDescription": "Counts demand data read requests that miss t= he L3 cache. Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -427,7 +423,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.L3_MISS_DEMAND_DATA_RD", - "PublicDescription": "For every cycle, increments by the number of= demand data read requests pending that are known to have missed the L3 cac= he. Note that this does not capture all elapsed cycles while requests are = outstanding - only cycles from when the requests were known by the requesti= ng core to have missed the L3 cache. Available PDIST counters: 0", + "PublicDescription": "For every cycle, increments by the number of= demand data read requests pending that are known to have missed the L3 cac= he. Note that this does not capture all elapsed cycles while requests are = outstanding - only cycles from when the requests were known by the requesti= ng core to have missed the L3 cache.", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/alderlake/other.json b/tools/pe= rf/pmu-events/arch/x86/alderlake/other.json index 5f64138edfe4..af46cde26b54 100644 --- a/tools/perf/pmu-events/arch/x86/alderlake/other.json +++ b/tools/perf/pmu-events/arch/x86/alderlake/other.json @@ -4,7 +4,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.HARDWARE", - "PublicDescription": "Count all other hardware assists or traps th= at are not necessarily architecturally exposed (through a software handler)= beyond FP; SSE-AVX mix and A/D assists who are counted by dedicated sub-ev= ents. This includes, but not limited to, assists at EXE or MEM uop writeba= ck like AVX* load/store/gather/scatter (non-FP GSSE-assist ) , assists gene= rated by ROB like PEBS and RTIT, Uncore trap, RAR (Remote Action Request) a= nd CET (Control flow Enforcement Technology) assists. the event also counts= for Machine Ordering count. Available PDIST counters: 0", + "PublicDescription": "Count all other hardware assists or traps th= at are not necessarily architecturally exposed (through a software handler)= beyond FP; SSE-AVX mix and A/D assists who are counted by dedicated sub-ev= ents. This includes, but not limited to, assists at EXE or MEM uop writeba= ck like AVX* load/store/gather/scatter (non-FP GSSE-assist ) , assists gene= rated by ROB like PEBS and RTIT, Uncore trap, RAR (Remote Action Request) a= nd CET (Control flow Enforcement Technology) assists. the event also counts= for Machine Ordering count.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -14,7 +14,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.PAGE_FAULT", - "PublicDescription": "ASSISTS.PAGE_FAULT Available PDIST counters:= 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -24,7 +23,6 @@ "Counter": "0,1,2,3", "EventCode": "0x28", "EventName": "CORE_POWER.LICENSE_1", - "PublicDescription": "CORE_POWER.LICENSE_1 Available PDIST counter= s: 0", "SampleAfterValue": "200003", "UMask": "0x2", "Unit": "cpu_core" @@ -34,7 +32,6 @@ "Counter": "0,1,2,3", "EventCode": "0x28", "EventName": "CORE_POWER.LICENSE_2", - "PublicDescription": "CORE_POWER.LICENSE_2 Available PDIST counter= s: 0", "SampleAfterValue": "200003", "UMask": "0x4", "Unit": "cpu_core" @@ -44,7 +41,6 @@ "Counter": "0,1,2,3", "EventCode": "0x28", "EventName": "CORE_POWER.LICENSE_3", - "PublicDescription": "CORE_POWER.LICENSE_3 Available PDIST counter= s: 0", "SampleAfterValue": "200003", "UMask": "0x8", "Unit": "cpu_core" @@ -113,7 +109,7 @@ "CounterMask": "1", "EventCode": "0x2d", "EventName": "XQ.FULL_CYCLES", - "PublicDescription": "number of cycles when the thread is active a= nd the uncore cannot take any further requests (for example prefetches, loa= ds or stores initiated by the Core that miss the L2 cache). Available PDIST= counters: 0", + "PublicDescription": "number of cycles when the thread is active a= nd the uncore cannot take any further requests (for example prefetches, loa= ds or stores initiated by the Core that miss the L2 cache).", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/alderlake/pipeline.json b/tools= /perf/pmu-events/arch/x86/alderlake/pipeline.json index 48ef2a8cc49a..33d1f39e441f 100644 --- a/tools/perf/pmu-events/arch/x86/alderlake/pipeline.json +++ b/tools/perf/pmu-events/arch/x86/alderlake/pipeline.json @@ -6,7 +6,6 @@ "Deprecated": "1", "EventCode": "0xb0", "EventName": "ARITH.DIVIDER_ACTIVE", - "PublicDescription": "This event is deprecated. Refer to new event= ARITH.DIV_ACTIVE Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x9", "Unit": "cpu_core" @@ -27,7 +26,7 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.DIV_ACTIVE", - "PublicDescription": "Counts cycles when divide unit is busy execu= ting divide or square root operations. Accounts for integer and floating-po= int operations. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when divide unit is busy execu= ting divide or square root operations. Accounts for integer and floating-po= int operations.", "SampleAfterValue": "1000003", "UMask": "0x9", "Unit": "cpu_core" @@ -57,7 +56,6 @@ "Deprecated": "1", "EventCode": "0xb0", "EventName": "ARITH.FP_DIVIDER_ACTIVE", - "PublicDescription": "This event is deprecated. Refer to new event= ARITH.FPDIV_ACTIVE Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -78,7 +76,6 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.IDIV_ACTIVE", - "PublicDescription": "This event counts the cycles the integer div= ider is busy. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -108,7 +105,6 @@ "Deprecated": "1", "EventCode": "0xb0", "EventName": "ARITH.INT_DIVIDER_ACTIVE", - "PublicDescription": "This event is deprecated. Refer to new event= ARITH.IDIV_ACTIVE Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -118,7 +114,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.ANY", - "PublicDescription": "Counts the number of occurrences where a mic= rocode assist is invoked by hardware. Examples include AD (page Access Dirt= y), FP and AVX related assists. Available PDIST counters: 0", + "PublicDescription": "Counts the number of occurrences where a mic= rocode assist is invoked by hardware. Examples include AD (page Access Dirt= y), FP and AVX related assists.", "SampleAfterValue": "100003", "UMask": "0x1b", "Unit": "cpu_core" @@ -549,7 +545,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C01", - "PublicDescription": "Counts core clocks when the thread is in the= C0.1 light-weight slower wakeup time but more power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions. Availab= le PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.1 light-weight slower wakeup time but more power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions.", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" @@ -559,7 +555,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C02", - "PublicDescription": "Counts core clocks when the thread is in the= C0.2 light-weight faster wakeup time but less power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions. Availab= le PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.2 light-weight faster wakeup time but less power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions.", "SampleAfterValue": "2000003", "UMask": "0x20", "Unit": "cpu_core" @@ -569,7 +565,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C0_WAIT", - "PublicDescription": "Counts core clocks when the thread is in the= C0.1 or C0.2 power saving optimized states (TPAUSE or UMWAIT instructions)= or running the PAUSE instruction. Available PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.1 or C0.2 power saving optimized states (TPAUSE or UMWAIT instructions)= or running the PAUSE instruction.", "SampleAfterValue": "2000003", "UMask": "0x70", "Unit": "cpu_core" @@ -597,7 +593,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.DISTRIBUTED", - "PublicDescription": "This event distributes cycle counts between = active hyperthreads, i.e., those in C0. A hyperthread becomes inactive whe= n it executes the HLT or MWAIT instructions. If all other hyperthreads are= inactive (or disabled or do not exist), all counts are attributed to this = hyperthread. To obtain the full count when the Core is active, sum the coun= ts from each hyperthread. Available PDIST counters: 0", + "PublicDescription": "This event distributes cycle counts between = active hyperthreads, i.e., those in C0. A hyperthread becomes inactive whe= n it executes the HLT or MWAIT instructions. If all other hyperthreads are= inactive (or disabled or do not exist), all counts are attributed to this = hyperthread. To obtain the full count when the Core is active, sum the coun= ts from each hyperthread.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -607,7 +603,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE", - "PublicDescription": "Counts Core crystal clock cycles when curren= t thread is unhalted and the other thread is halted. Available PDIST counte= rs: 0", + "PublicDescription": "Counts Core crystal clock cycles when curren= t thread is unhalted and the other thread is halted.", "SampleAfterValue": "25003", "UMask": "0x2", "Unit": "cpu_core" @@ -617,7 +613,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.PAUSE", - "PublicDescription": "CPU_CLK_UNHALTED.PAUSE Available PDIST count= ers: 0", "SampleAfterValue": "2000003", "UMask": "0x40", "Unit": "cpu_core" @@ -629,7 +624,6 @@ "EdgeDetect": "1", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.PAUSE_INST", - "PublicDescription": "CPU_CLK_UNHALTED.PAUSE_INST Available PDIST = counters: 0", "SampleAfterValue": "2000003", "UMask": "0x40", "Unit": "cpu_core" @@ -649,7 +643,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.REF_DISTRIBUTED", - "PublicDescription": "This event distributes Core crystal clock cy= cle counts between active hyperthreads, i.e., those in C0 sleep-state. A hy= perthread becomes inactive when it executes the HLT or MWAIT instructions. = If one thread is active in a core, all counts are attributed to this hypert= hread. To obtain the full count when the Core is active, sum the counts fro= m each hyperthread. Available PDIST counters: 0", + "PublicDescription": "This event distributes Core crystal clock cy= cle counts between active hyperthreads, i.e., those in C0 sleep-state. A hy= perthread becomes inactive when it executes the HLT or MWAIT instructions. = If one thread is active in a core, all counts are attributed to this hypert= hread. To obtain the full count when the Core is active, sum the counts fro= m each hyperthread.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -687,7 +681,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.REF_TSC_P", - "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the four (eight when Hyperthr= eading is disabled) programmable counters available for other events. Note:= On all current platforms this event stops counting during 'throttling (TM)= ' states duty off periods the processor is 'halted'. The counter update is= done at a lower clock rate then the core clock the overflow status bit for= this counter may appear 'sticky'. After the counter has overflowed and so= ftware clears the overflow status bit and resets the counter to less than M= AX. The reset value to the counter is not clocked immediately so the overfl= ow status bit will flip 'high (1)' and generate another PMI (if enabled) af= ter which the reset value gets clocked into the counter. Therefore, softwar= e will get the interrupt, read the overflow status bit '1 for bit 34 while = the counter value is less than MAX. Software should ignore this case. Avail= able PDIST counters: 0", + "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the four (eight when Hyperthr= eading is disabled) programmable counters available for other events. Note:= On all current platforms this event stops counting during 'throttling (TM)= ' states duty off periods the processor is 'halted'. The counter update is= done at a lower clock rate then the core clock the overflow status bit for= this counter may appear 'sticky'. After the counter has overflowed and so= ftware clears the overflow status bit and resets the counter to less than M= AX. The reset value to the counter is not clocked immediately so the overfl= ow status bit will flip 'high (1)' and generate another PMI (if enabled) af= ter which the reset value gets clocked into the counter. Therefore, softwar= e will get the interrupt, read the overflow status bit '1 for bit 34 while = the counter value is less than MAX. Software should ignore this case.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -724,7 +718,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.THREAD_P", - "PublicDescription": "This is an architectural event that counts t= he number of thread cycles while the thread is not in a halt state. The thr= ead enters the halt state when it is running the HLT instruction. The core = frequency may change from time to time due to power or thermal throttling. = For this reason, this event may have a changing ratio with regards to wall = clock time. Available PDIST counters: 0", + "PublicDescription": "This is an architectural event that counts t= he number of thread cycles while the thread is not in a halt state. The thr= ead enters the halt state when it is running the HLT instruction. The core = frequency may change from time to time due to power or thermal throttling. = For this reason, this event may have a changing ratio with regards to wall = clock time.", "SampleAfterValue": "2000003", "Unit": "cpu_core" }, @@ -734,7 +728,6 @@ "CounterMask": "8", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS", - "PublicDescription": "Cycles while L1 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -745,7 +738,6 @@ "CounterMask": "1", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS", - "PublicDescription": "Cycles while L2 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -756,7 +748,6 @@ "CounterMask": "16", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY", - "PublicDescription": "Cycles while memory subsystem has an outstan= ding load. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -767,7 +758,6 @@ "CounterMask": "12", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS", - "PublicDescription": "Execution stalls while L1 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0xc", "Unit": "cpu_core" @@ -778,7 +768,6 @@ "CounterMask": "5", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS", - "PublicDescription": "Execution stalls while L2 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x5", "Unit": "cpu_core" @@ -789,7 +778,6 @@ "CounterMask": "4", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL", - "PublicDescription": "Total execution stalls. Available PDIST coun= ters: 0", "SampleAfterValue": "1000003", "UMask": "0x4", "Unit": "cpu_core" @@ -799,7 +787,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.1_PORTS_UTIL", - "PublicDescription": "Counts cycles during which a total of 1 uop = was executed on all ports and Reservation Station (RS) was not empty. Avail= able PDIST counters: 0", + "PublicDescription": "Counts cycles during which a total of 1 uop = was executed on all ports and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -809,7 +797,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.2_3_PORTS_UTIL", - "PublicDescription": "Cycles total of 2 or 3 uops are executed on = all ports and Reservation Station (RS) was not empty. Available PDIST count= ers: 0", "SampleAfterValue": "2000003", "UMask": "0xc", "Unit": "cpu_core" @@ -819,7 +806,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.2_PORTS_UTIL", - "PublicDescription": "Counts cycles during which a total of 2 uops= were executed on all ports and Reservation Station (RS) was not empty. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts cycles during which a total of 2 uops= were executed on all ports and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -829,7 +816,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.3_PORTS_UTIL", - "PublicDescription": "Cycles total of 3 uops are executed on all p= orts and Reservation Station (RS) was not empty. Available PDIST counters: = 0", + "PublicDescription": "Cycles total of 3 uops are executed on all p= orts and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -839,7 +826,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.4_PORTS_UTIL", - "PublicDescription": "Cycles total of 4 uops are executed on all p= orts and Reservation Station (RS) was not empty. Available PDIST counters: = 0", + "PublicDescription": "Cycles total of 4 uops are executed on all p= orts and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" @@ -850,7 +837,6 @@ "CounterMask": "5", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.BOUND_ON_LOADS", - "PublicDescription": "Execution stalls while memory subsystem has = an outstanding load. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x21", "Unit": "cpu_core" @@ -861,7 +847,7 @@ "CounterMask": "2", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.BOUND_ON_STORES", - "PublicDescription": "Counts cycles where the Store Buffer was ful= l and no loads caused an execution stall. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where the Store Buffer was ful= l and no loads caused an execution stall.", "SampleAfterValue": "1000003", "UMask": "0x40", "Unit": "cpu_core" @@ -871,7 +857,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.EXE_BOUND_0_PORTS", - "PublicDescription": "Number of cycles total of 0 uops executed on= all ports, Reservation Station (RS) was not empty, the Store Buffer (SB) w= as not full and there was no outstanding load. Available PDIST counters: 0", + "PublicDescription": "Number of cycles total of 0 uops executed on= all ports, Reservation Station (RS) was not empty, the Store Buffer (SB) w= as not full and there was no outstanding load.", "SampleAfterValue": "1000003", "UMask": "0x80", "Unit": "cpu_core" @@ -881,7 +867,7 @@ "Counter": "0,1,2,3", "EventCode": "0x75", "EventName": "INST_DECODED.DECODERS", - "PublicDescription": "Number of decoders utilized in a cycle when = the MITE (legacy decode pipeline) fetches instructions. Available PDIST cou= nters: 0", + "PublicDescription": "Number of decoders utilized in a cycle when = the MITE (legacy decode pipeline) fetches instructions.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -927,7 +913,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.MACRO_FUSED", - "PublicDescription": "INST_RETIRED.MACRO_FUSED Available PDIST cou= nters: 0", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" @@ -937,7 +922,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.NOP", - "PublicDescription": "Counts all retired NOP or ENDBR32/64 instruc= tions Available PDIST counters: 0", + "PublicDescription": "Counts all retired NOP or ENDBR32/64 instruc= tions", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -956,7 +941,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.REP_ITERATION", - "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent. Available PDIST counters: 0", + "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -968,7 +953,7 @@ "EdgeDetect": "1", "EventCode": "0xad", "EventName": "INT_MISC.CLEARS_COUNT", - "PublicDescription": "Counts the number of speculative clears due = to any type of branch misprediction or machine clears Available PDIST count= ers: 0", + "PublicDescription": "Counts the number of speculative clears due = to any type of branch misprediction or machine clears", "SampleAfterValue": "500009", "UMask": "0x1", "Unit": "cpu_core" @@ -978,7 +963,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.CLEAR_RESTEER_CYCLES", - "PublicDescription": "Cycles after recovery from a branch mispredi= ction or machine clear till the first uop is issued from the resteered path= . Available PDIST counters: 0", + "PublicDescription": "Cycles after recovery from a branch mispredi= ction or machine clear till the first uop is issued from the resteered path= .", "SampleAfterValue": "500009", "UMask": "0x80", "Unit": "cpu_core" @@ -988,7 +973,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.RECOVERY_CYCLES", - "PublicDescription": "Counts core cycles when the Resource allocat= or was stalled due to recovery from an earlier branch misprediction or mach= ine clear event. Available PDIST counters: 0", + "PublicDescription": "Counts core cycles when the Resource allocat= or was stalled due to recovery from an earlier branch misprediction or mach= ine clear event.", "SampleAfterValue": "500009", "UMask": "0x1", "Unit": "cpu_core" @@ -1000,7 +985,6 @@ "EventName": "INT_MISC.UNKNOWN_BRANCH_CYCLES", "MSRIndex": "0x3F7", "MSRValue": "0x7", - "PublicDescription": "Bubble cycles of BAClear (Unknown Branch). A= vailable PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x40", "Unit": "cpu_core" @@ -1010,7 +994,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.UOP_DROPPING", - "PublicDescription": "Estimated number of Top-down Microarchitectu= re Analysis slots that got dropped due to non front-end reasons Available P= DIST counters: 0", + "PublicDescription": "Estimated number of Top-down Microarchitectu= re Analysis slots that got dropped due to non front-end reasons", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -1020,7 +1004,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.128BIT", - "PublicDescription": "INT_VEC_RETIRED.128BIT Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0x13", "Unit": "cpu_core" @@ -1030,7 +1013,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.256BIT", - "PublicDescription": "INT_VEC_RETIRED.256BIT Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0xac", "Unit": "cpu_core" @@ -1040,7 +1022,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.ADD_128", - "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 128-bit vector instructions. Available PDIST counters: 0= ", + "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 128-bit vector instructions.", "SampleAfterValue": "1000003", "UMask": "0x3", "Unit": "cpu_core" @@ -1050,7 +1032,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.ADD_256", - "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 256-bit vector instructions. Available PDIST counters: 0= ", + "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 256-bit vector instructions.", "SampleAfterValue": "1000003", "UMask": "0xc", "Unit": "cpu_core" @@ -1060,7 +1042,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.MUL_256", - "PublicDescription": "INT_VEC_RETIRED.MUL_256 Available PDIST coun= ters: 0", "SampleAfterValue": "1000003", "UMask": "0x80", "Unit": "cpu_core" @@ -1070,7 +1051,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.SHUFFLES", - "PublicDescription": "INT_VEC_RETIRED.SHUFFLES Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x40", "Unit": "cpu_core" @@ -1080,7 +1060,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.VNNI_128", - "PublicDescription": "INT_VEC_RETIRED.VNNI_128 Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -1090,7 +1069,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.VNNI_256", - "PublicDescription": "INT_VEC_RETIRED.VNNI_256 Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x20", "Unit": "cpu_core" @@ -1119,7 +1097,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.ADDRESS_ALIAS", - "PublicDescription": "Counts the number of times a load got blocke= d due to false dependencies in MOB due to partial compare on address. Avail= able PDIST counters: 0", + "PublicDescription": "Counts the number of times a load got blocke= d due to false dependencies in MOB due to partial compare on address.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -1138,7 +1116,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.NO_SR", - "PublicDescription": "Counts the number of times that split load o= perations are temporarily blocked because all resources for handling the sp= lit accesses are in use. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times that split load o= perations are temporarily blocked because all resources for handling the sp= lit accesses are in use.", "SampleAfterValue": "100003", "UMask": "0x88", "Unit": "cpu_core" @@ -1148,7 +1126,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.STORE_FORWARD", - "PublicDescription": "Counts the number of times where store forwa= rding was prevented for a load operation. The most common case is a load bl= ocked due to the address of memory access (partially) overlapping with a pr= eceding uncompleted store. Note: See the table of not supported store forwa= rds in the Optimization Guide. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times where store forwa= rding was prevented for a load operation. The most common case is a load bl= ocked due to the address of memory access (partially) overlapping with a pr= eceding uncompleted store. Note: See the table of not supported store forwa= rds in the Optimization Guide.", "SampleAfterValue": "100003", "UMask": "0x82", "Unit": "cpu_core" @@ -1158,7 +1136,7 @@ "Counter": "0,1,2,3", "EventCode": "0x4c", "EventName": "LOAD_HIT_PREFETCH.SWPF", - "PublicDescription": "Counts all software-prefetch load dispatches= that hit the fill buffer (FB) allocated for the software prefetch. It can = also be incremented by some lock instructions. So it should only be used wi= th profiling so that the locks can be excluded by ASM (Assembly File) inspe= ction of the nearby instructions. Available PDIST counters: 0", + "PublicDescription": "Counts all software-prefetch load dispatches= that hit the fill buffer (FB) allocated for the software prefetch. It can = also be incremented by some lock instructions. So it should only be used wi= th profiling so that the locks can be excluded by ASM (Assembly File) inspe= ction of the nearby instructions.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -1169,7 +1147,7 @@ "CounterMask": "1", "EventCode": "0xa8", "EventName": "LSD.CYCLES_ACTIVE", - "PublicDescription": "Counts the cycles when at least one uop is d= elivered by the LSD (Loop-stream detector). Available PDIST counters: 0", + "PublicDescription": "Counts the cycles when at least one uop is d= elivered by the LSD (Loop-stream detector).", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1180,7 +1158,7 @@ "CounterMask": "6", "EventCode": "0xa8", "EventName": "LSD.CYCLES_OK", - "PublicDescription": "Counts the cycles when optimal number of uop= s is delivered by the LSD (Loop-stream detector). Available PDIST counters:= 0", + "PublicDescription": "Counts the cycles when optimal number of uop= s is delivered by the LSD (Loop-stream detector).", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1190,7 +1168,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa8", "EventName": "LSD.UOPS", - "PublicDescription": "Counts the number of uops delivered to the b= ack-end by the LSD(Loop Stream Detector). Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops delivered to the b= ack-end by the LSD(Loop Stream Detector).", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1202,7 +1180,7 @@ "EdgeDetect": "1", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.COUNT", - "PublicDescription": "Counts the number of machine clears (nukes) = of any type. Available PDIST counters: 0", + "PublicDescription": "Counts the number of machine clears (nukes) = of any type.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -1258,7 +1236,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.SMC", - "PublicDescription": "Counts self-modifying code (SMC) detected, w= hich causes a machine clear. Available PDIST counters: 0", + "PublicDescription": "Counts self-modifying code (SMC) detected, w= hich causes a machine clear.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -1268,7 +1246,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe0", "EventName": "MISC2_RETIRED.LFENCE", - "PublicDescription": "number of LFENCE retired instructions Availa= ble PDIST counters: 0", + "PublicDescription": "number of LFENCE retired instructions", "SampleAfterValue": "400009", "UMask": "0x20", "Unit": "cpu_core" @@ -1288,7 +1266,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcc", "EventName": "MISC_RETIRED.LBR_INSERTS", - "PublicDescription": "Increments when an entry is added to the Las= t Branch Record (LBR) array (or removed from the array in case of RETURNs i= n call stack mode). The event requires LBR enable via IA32_DEBUGCTL MSR and= branch type selection via MSR_LBR_SELECT. Available PDIST counters: 0", + "PublicDescription": "Increments when an entry is added to the Las= t Branch Record (LBR) array (or removed from the array in case of RETURNs i= n call stack mode). The event requires LBR enable via IA32_DEBUGCTL MSR and= branch type selection via MSR_LBR_SELECT.", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -1298,7 +1276,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa2", "EventName": "RESOURCE_STALLS.SB", - "PublicDescription": "Counts allocation stall cycles caused by the= store buffer (SB) being full. This counts cycles that the pipeline back-en= d blocked uop delivery from the front-end. Available PDIST counters: 0", + "PublicDescription": "Counts allocation stall cycles caused by the= store buffer (SB) being full. This counts cycles that the pipeline back-en= d blocked uop delivery from the front-end.", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -1308,7 +1286,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa2", "EventName": "RESOURCE_STALLS.SCOREBOARD", - "PublicDescription": "Counts cycles where the pipeline is stalled = due to serializing operations. Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -1318,7 +1295,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa5", "EventName": "RS.EMPTY", - "PublicDescription": "Counts cycles during which the reservation s= tation (RS) is empty for this logical processor. This is usually caused whe= n the front-end pipeline runs into starvation periods (e.g. branch mispredi= ctions or i-cache misses) Available PDIST counters: 0", + "PublicDescription": "Counts cycles during which the reservation s= tation (RS) is empty for this logical processor. This is usually caused whe= n the front-end pipeline runs into starvation periods (e.g. branch mispredi= ctions or i-cache misses)", "SampleAfterValue": "1000003", "UMask": "0x7", "Unit": "cpu_core" @@ -1331,7 +1308,7 @@ "EventCode": "0xa5", "EventName": "RS.EMPTY_COUNT", "Invert": "1", - "PublicDescription": "Counts end of periods where the Reservation = Station (RS) was empty. Could be useful to closely sample on front-end late= ncy issues (see the FRONTEND_RETIRED event of designated precise events) Av= ailable PDIST counters: 0", + "PublicDescription": "Counts end of periods where the Reservation = Station (RS) was empty. Could be useful to closely sample on front-end late= ncy issues (see the FRONTEND_RETIRED event of designated precise events)", "SampleAfterValue": "100003", "UMask": "0x7", "Unit": "cpu_core" @@ -1341,7 +1318,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa5", "EventName": "RS.EMPTY_RESOURCE", - "PublicDescription": "Cycles when Reservation Station (RS) is empt= y due to a resource in the back-end Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1355,7 +1331,6 @@ "EventCode": "0xa5", "EventName": "RS_EMPTY.COUNT", "Invert": "1", - "PublicDescription": "This event is deprecated. Refer to new event= RS.EMPTY_COUNT Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x7", "Unit": "cpu_core" @@ -1366,7 +1341,6 @@ "Deprecated": "1", "EventCode": "0xa5", "EventName": "RS_EMPTY.CYCLES", - "PublicDescription": "This event is deprecated. Refer to new event= RS.EMPTY Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x7", "Unit": "cpu_core" @@ -1395,7 +1369,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.BACKEND_BOUND_SLOTS", - "PublicDescription": "Number of slots in TMA method where no micro= -operations were being issued from front-end to back-end of the machine due= to lack of back-end resources. Available PDIST counters: 0", + "PublicDescription": "Number of slots in TMA method where no micro= -operations were being issued from front-end to back-end of the machine due= to lack of back-end resources.", "SampleAfterValue": "10000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1405,7 +1379,7 @@ "Counter": "0", "EventCode": "0xa4", "EventName": "TOPDOWN.BAD_SPEC_SLOTS", - "PublicDescription": "Number of slots of TMA method that were wast= ed due to incorrect speculation. It covers all types of control-flow or dat= a-related mis-speculations. Available PDIST counters: 0", + "PublicDescription": "Number of slots of TMA method that were wast= ed due to incorrect speculation. It covers all types of control-flow or dat= a-related mis-speculations.", "SampleAfterValue": "10000003", "UMask": "0x4", "Unit": "cpu_core" @@ -1415,7 +1389,7 @@ "Counter": "0", "EventCode": "0xa4", "EventName": "TOPDOWN.BR_MISPREDICT_SLOTS", - "PublicDescription": "Number of TMA slots that were wasted due to = incorrect speculation by (any type of) branch mispredictions. This event es= timates number of speculative operations that were issued but not retired a= s well as the out-of-order engine recovery past a branch misprediction. Ava= ilable PDIST counters: 0", + "PublicDescription": "Number of TMA slots that were wasted due to = incorrect speculation by (any type of) branch mispredictions. This event es= timates number of speculative operations that were issued but not retired a= s well as the out-of-order engine recovery past a branch misprediction.", "SampleAfterValue": "10000003", "UMask": "0x8", "Unit": "cpu_core" @@ -1425,7 +1399,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.MEMORY_BOUND_SLOTS", - "PublicDescription": "TOPDOWN.MEMORY_BOUND_SLOTS Available PDIST c= ounters: 0", "SampleAfterValue": "10000003", "UMask": "0x10", "Unit": "cpu_core" @@ -1444,7 +1417,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.SLOTS_P", - "PublicDescription": "Counts the number of available slots for an = unhalted logical processor. The event increments by machine-width of the na= rrowest pipeline as employed by the Top-down Microarchitecture Analysis met= hod. The count is distributed among unhalted logical processors (hyper-thre= ads) who share the same physical core. Available PDIST counters: 0", + "PublicDescription": "Counts the number of available slots for an = unhalted logical processor. The event increments by machine-width of the na= rrowest pipeline as employed by the Top-down Microarchitecture Analysis met= hod. The count is distributed among unhalted logical processors (hyper-thre= ads) who share the same physical core.", "SampleAfterValue": "10000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1661,7 +1634,6 @@ "Counter": "0,1,2,3", "EventCode": "0x76", "EventName": "UOPS_DECODED.DEC0_UOPS", - "PublicDescription": "UOPS_DECODED.DEC0_UOPS Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1671,7 +1643,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_0", - "PublicDescription": "Number of uops dispatch to execution port 0= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 0= .", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1681,7 +1653,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_1", - "PublicDescription": "Number of uops dispatch to execution port 1= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 1= .", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1691,7 +1663,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_2_3_10", - "PublicDescription": "Number of uops dispatch to execution ports 2= , 3 and 10 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 2= , 3 and 10", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -1701,7 +1673,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_4_9", - "PublicDescription": "Number of uops dispatch to execution ports 4= and 9 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 4= and 9", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" @@ -1711,7 +1683,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_5_11", - "PublicDescription": "Number of uops dispatch to execution ports 5= and 11 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 5= and 11", "SampleAfterValue": "2000003", "UMask": "0x20", "Unit": "cpu_core" @@ -1721,7 +1693,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_6", - "PublicDescription": "Number of uops dispatch to execution port 6= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 6= .", "SampleAfterValue": "2000003", "UMask": "0x40", "Unit": "cpu_core" @@ -1731,7 +1703,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_7_8", - "PublicDescription": "Number of uops dispatch to execution ports = 7 and 8. Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports = 7 and 8.", "SampleAfterValue": "2000003", "UMask": "0x80", "Unit": "cpu_core" @@ -1742,7 +1714,7 @@ "CounterMask": "1", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1", - "PublicDescription": "Counts cycles when at least 1 micro-op is ex= ecuted from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 1 micro-op is ex= ecuted from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1753,7 +1725,7 @@ "CounterMask": "2", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2", - "PublicDescription": "Counts cycles when at least 2 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 2 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1764,7 +1736,7 @@ "CounterMask": "3", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3", - "PublicDescription": "Counts cycles when at least 3 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 3 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1775,7 +1747,7 @@ "CounterMask": "4", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4", - "PublicDescription": "Counts cycles when at least 4 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 4 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1786,7 +1758,7 @@ "CounterMask": "1", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_1", - "PublicDescription": "Cycles where at least 1 uop was executed per= -thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 1 uop was executed per= -thread.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1797,7 +1769,7 @@ "CounterMask": "2", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_2", - "PublicDescription": "Cycles where at least 2 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 2 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1808,7 +1780,7 @@ "CounterMask": "3", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_3", - "PublicDescription": "Cycles where at least 3 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 3 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1819,7 +1791,7 @@ "CounterMask": "4", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_4", - "PublicDescription": "Cycles where at least 4 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 4 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1831,7 +1803,7 @@ "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.STALLS", "Invert": "1", - "PublicDescription": "Counts cycles during which no uops were disp= atched from the Reservation Station (RS) per thread. Available PDIST counte= rs: 0", + "PublicDescription": "Counts cycles during which no uops were disp= atched from the Reservation Station (RS) per thread.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1844,7 +1816,6 @@ "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.STALL_CYCLES", "Invert": "1", - "PublicDescription": "This event is deprecated. Refer to new event= UOPS_EXECUTED.STALLS Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1854,7 +1825,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.THREAD", - "PublicDescription": "Counts the number of uops to be executed per= -thread each cycle. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1864,7 +1834,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.X87", - "PublicDescription": "Counts the number of x87 uops executed. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts the number of x87 uops executed.", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" @@ -1883,7 +1853,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xae", "EventName": "UOPS_ISSUED.ANY", - "PublicDescription": "Counts the number of uops that the Resource = Allocation Table (RAT) issues to the Reservation Station (RS). Available PD= IST counters: 0", + "PublicDescription": "Counts the number of uops that the Resource = Allocation Table (RAT) issues to the Reservation Station (RS).", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1894,7 +1864,6 @@ "CounterMask": "1", "EventCode": "0xae", "EventName": "UOPS_ISSUED.CYCLES", - "PublicDescription": "UOPS_ISSUED.CYCLES Available PDIST counters:= 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1913,7 +1882,7 @@ "CounterMask": "1", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.CYCLES", - "PublicDescription": "Counts cycles where at least one uop has ret= ired. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where at least one uop has ret= ired.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1923,7 +1892,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.HEAVY", - "PublicDescription": "Counts the number of retired micro-operation= s (uops) except the last uop of each instruction. An instruction that is de= coded into less than two uops does not contribute to the count. Available P= DIST counters: 0", + "PublicDescription": "Counts the number of retired micro-operation= s (uops) except the last uop of each instruction. An instruction that is de= coded into less than two uops does not contribute to the count.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1954,7 +1923,6 @@ "EventName": "UOPS_RETIRED.MS", "MSRIndex": "0x3F7", "MSRValue": "0x8", - "PublicDescription": "UOPS_RETIRED.MS Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -1964,7 +1932,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.SLOTS", - "PublicDescription": "Counts the retirement slots used each cycle.= Available PDIST counters: 0", + "PublicDescription": "Counts the retirement slots used each cycle.= ", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1976,7 +1944,7 @@ "EventCode": "0xc2", "EventName": "UOPS_RETIRED.STALLS", "Invert": "1", - "PublicDescription": "This event counts cycles without actually re= tired uops. Available PDIST counters: 0", + "PublicDescription": "This event counts cycles without actually re= tired uops.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1989,7 +1957,6 @@ "EventCode": "0xc2", "EventName": "UOPS_RETIRED.STALL_CYCLES", "Invert": "1", - "PublicDescription": "This event is deprecated. Refer to new event= UOPS_RETIRED.STALLS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/alderlake/uncore-interconnect.j= son b/tools/perf/pmu-events/arch/x86/alderlake/uncore-interconnect.json index 7c0779c74154..b5604c7534e1 100644 --- a/tools/perf/pmu-events/arch/x86/alderlake/uncore-interconnect.json +++ b/tools/perf/pmu-events/arch/x86/alderlake/uncore-interconnect.json @@ -65,7 +65,6 @@ "Counter": "0,1", "EventCode": "0x81", "EventName": "UNC_ARB_REQ_TRK_REQUEST.DRD", - "Experimental": "1", "PerPkg": "1", "UMask": "0x2", "Unit": "ARB" @@ -103,7 +102,6 @@ "Counter": "0,1", "EventCode": "0x81", "EventName": "UNC_ARB_TRK_REQUESTS.RD", - "Experimental": "1", "PerPkg": "1", "UMask": "0x2", "Unit": "ARB" diff --git a/tools/perf/pmu-events/arch/x86/alderlake/virtual-memory.json b= /tools/perf/pmu-events/arch/x86/alderlake/virtual-memory.json index ffbbd08acc68..132ce48af6d9 100644 --- a/tools/perf/pmu-events/arch/x86/alderlake/virtual-memory.json +++ b/tools/perf/pmu-events/arch/x86/alderlake/virtual-memory.json @@ -4,7 +4,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.STLB_HIT", - "PublicDescription": "Counts loads that miss the DTLB (Data TLB) a= nd hit the STLB (Second level TLB). Available PDIST counters: 0", + "PublicDescription": "Counts loads that miss the DTLB (Data TLB) a= nd hit the STLB (Second level TLB).", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -15,7 +15,7 @@ "CounterMask": "1", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a demand load. Available PDIST cou= nters: 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a demand load.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -35,7 +35,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data loads. This implies it missed in the DTLB and furth= er levels of TLB. The page walk can end with or without a fault. Available = PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data loads. This implies it missed in the DTLB and furth= er levels of TLB. The page walk can end with or without a fault.", "SampleAfterValue": "100003", "UMask": "0xe", "Unit": "cpu_core" @@ -45,7 +45,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_1G", - "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -55,7 +55,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data loads. This implies address translations missed in the= DTLB and further levels of TLB. The page walk can end with or without a fa= ult. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data loads. This implies address translations missed in the= DTLB and further levels of TLB. The page walk can end with or without a fa= ult.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -65,7 +65,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -75,7 +75,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for a demand load in the PMH (Page Miss Handler) each cycle. Available PDIS= T counters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for a demand load in the PMH (Page Miss Handler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -85,7 +85,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.STLB_HIT", - "PublicDescription": "Counts stores that miss the DTLB (Data TLB) = and hit the STLB (2nd Level TLB). Available PDIST counters: 0", + "PublicDescription": "Counts stores that miss the DTLB (Data TLB) = and hit the STLB (2nd Level TLB).", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -96,7 +96,7 @@ "CounterMask": "1", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a store. Available PDIST counters:= 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a store.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -116,7 +116,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data stores. This implies it missed in the DTLB and furt= her levels of TLB. The page walk can end with or without a fault. Available= PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data stores. This implies it missed in the DTLB and furt= her levels of TLB. The page walk can end with or without a fault.", "SampleAfterValue": "100003", "UMask": "0xe", "Unit": "cpu_core" @@ -126,7 +126,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_1G", - "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t.", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -136,7 +136,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data stores. This implies address translations missed in th= e DTLB and further levels of TLB. The page walk can end with or without a f= ault. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data stores. This implies address translations missed in th= e DTLB and further levels of TLB. The page walk can end with or without a f= ault.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -146,7 +146,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t.", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -156,7 +156,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for a store in the PMH (Page Miss Handler) each cycle. Available PDIST coun= ters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for a store in the PMH (Page Miss Handler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -184,7 +184,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.STLB_HIT", - "PublicDescription": "Counts instruction fetch requests that miss = the ITLB (Instruction TLB) and hit the STLB (Second-level TLB). Available P= DIST counters: 0", + "PublicDescription": "Counts instruction fetch requests that miss = the ITLB (Instruction TLB) and hit the STLB (Second-level TLB).", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -195,7 +195,7 @@ "CounterMask": "1", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a code (instruction fetch) request= . Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a code (instruction fetch) request= .", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -215,7 +215,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes)= caused by a code fetch. This implies it missed in the ITLB (Instruction TL= B) and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes)= caused by a code fetch. This implies it missed in the ITLB (Instruction TL= B) and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0xe", "Unit": "cpu_core" @@ -225,7 +225,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M page size= s) caused by a code fetch. This implies it missed in the ITLB (Instruction = TLB) and further levels of TLB. The page walk can end with or without a fau= lt. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M page size= s) caused by a code fetch. This implies it missed in the ITLB (Instruction = TLB) and further levels of TLB. The page walk can end with or without a fau= lt.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -235,7 +235,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K page sizes) = caused by a code fetch. This implies it missed in the ITLB (Instruction TLB= ) and further levels of TLB. The page walk can end with or without a fault.= Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K page sizes) = caused by a code fetch. This implies it missed in the ITLB (Instruction TLB= ) and further levels of TLB. The page walk can end with or without a fault.= ", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -245,7 +245,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for an outstanding code (instruction fetch) request in the PMH (Page Miss H= andler) each cycle. Available PDIST counters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for an outstanding code (instruction fetch) request in the PMH (Page Miss H= andler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/alderlaken/adln-metrics.json b/= tools/perf/pmu-events/arch/x86/alderlaken/adln-metrics.json index ce93648043ef..0f72c9192df6 100644 --- a/tools/perf/pmu-events/arch/x86/alderlaken/adln-metrics.json +++ b/tools/perf/pmu-events/arch/x86/alderlaken/adln-metrics.json @@ -1,56 +1,56 @@ [ { "BriefDescription": "C10 residency percent per package", - "MetricExpr": "cstate_pkg@c10\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c10\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C10_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C8 residency percent per package", - "MetricExpr": "cstate_pkg@c8\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c8\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C8_Pkg_Residency", "ScaleUnit": "100%" @@ -460,12 +460,12 @@ }, { "BriefDescription": "Average CPU Utilization", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricName": "tma_info_system_cpu_utilization" }, { "BriefDescription": "Fraction of cycles spent in Kernel mode", - "MetricExpr": "cpu@CPU_CLK_UNHALTED.CORE_P@k / CPU_CLK_UNHALTED.CO= RE", + "MetricExpr": "CPU_CLK_UNHALTED.CORE_P:k / CPU_CLK_UNHALTED.CORE", "MetricGroup": "Summary", "MetricName": "tma_info_system_kernel_utilization" }, diff --git a/tools/perf/pmu-events/arch/x86/alderlaken/uncore-interconnect.= json b/tools/perf/pmu-events/arch/x86/alderlaken/uncore-interconnect.json index 7c0779c74154..b5604c7534e1 100644 --- a/tools/perf/pmu-events/arch/x86/alderlaken/uncore-interconnect.json +++ b/tools/perf/pmu-events/arch/x86/alderlaken/uncore-interconnect.json @@ -65,7 +65,6 @@ "Counter": "0,1", "EventCode": "0x81", "EventName": "UNC_ARB_REQ_TRK_REQUEST.DRD", - "Experimental": "1", "PerPkg": "1", "UMask": "0x2", "Unit": "ARB" @@ -103,7 +102,6 @@ "Counter": "0,1", "EventCode": "0x81", "EventName": "UNC_ARB_TRK_REQUESTS.RD", - "Experimental": "1", "PerPkg": "1", "UMask": "0x2", "Unit": "ARB" diff --git a/tools/perf/pmu-events/arch/x86/mapfile.csv b/tools/perf/pmu-ev= ents/arch/x86/mapfile.csv index 354ce241500b..d0a17905c74e 100644 --- a/tools/perf/pmu-events/arch/x86/mapfile.csv +++ b/tools/perf/pmu-events/arch/x86/mapfile.csv @@ -1,6 +1,6 @@ Family-model,Version,Filename,EventType -GenuineIntel-6-(97|9A|B7|BA|BF),v1.31,alderlake,core -GenuineIntel-6-BE,v1.31,alderlaken,core +GenuineIntel-6-(97|9A|B7|BA|BF),v1.33,alderlake,core +GenuineIntel-6-BE,v1.33,alderlaken,core GenuineIntel-6-C[56],v1.09,arrowlake,core GenuineIntel-6-(1C|26|27|35|36),v5,bonnell,core GenuineIntel-6-(3D|47),v30,broadwell,core --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pl1-f202.google.com (mail-pl1-f202.google.com [209.85.214.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 4475213C3F6 for ; Mon, 18 Aug 2025 19:05:27 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.202 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543933; cv=none; b=fwws+lzSKMuBdh7ys8bGCncurzFNDBtCwp1NewO0ycXVbLZrZNJ87y5xHUuIkWwX9kO6bbehtTWhHLvt+t13vw3M2p3+Wbyqgez3KEulIPj9Eza65xBH3Cg7qWRe5PjY3QLM3V6YHeXfGoO20XdndJwKCcKqGYAIoTbV7FjwMdU= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543933; c=relaxed/simple; bh=DxQ0v2TSM1NqkYBeihUn2Yf12rAjcztxdgjh9xwdV3A=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=gTMSnRGZeuMkcRjTgZUccwTiYo0zdmMLnQdkdr+uLNYKll65PRfDpG1SIWLxwx4e/mm2JxlxRZYaYRXiK1VtflEajBtRnz7s7GCGYQzzCIgu5XOoLlbeB4G4/EVIaDInB51PVJhS1HU3PqMwFMDuMLbuc+05xuaWulm+Pyl2ksQ= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=n7IDoIky; arc=none smtp.client-ip=209.85.214.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="n7IDoIky" Received: by mail-pl1-f202.google.com with SMTP id d9443c01a7336-24457fe8844so51604275ad.0 for ; Mon, 18 Aug 2025 12:05:27 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543926; x=1756148726; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=bUE0KLqqqpCnYGAYYSp4QCNucfbWpvhcBRCQ0ZWWQjE=; b=n7IDoIkyCxFByiSG38DLj82E1KqUKMcMzz7TMFVK7q8itfx6mGu/829r/CvN/0Vgwc 7ujpPFnQypMbweJDA7n9lCn0NxWTuPA0D6K+okkjBi3NBj8Ye3XOr93oLHxdw8OxXhKN 2fbunV8Qq0lKQhUquL3hv8hjHuYsbZ1hMWys+f28NHoumlarF74qtsJYED5fYiGPPPiQ 5Q2Ix2azFd39g+Pmbp9rrnCcr4F5Uz87Dif+zam0euX53zy7TPoJ9ZjAS7CYuOCsV5Nl vfbdl0zuZ/1ioQgZertHlT2ATPljsUbeXXtYoC2rlycczYHWzDtQZp2EO+20mE1saU2a vLzg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543926; x=1756148726; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=bUE0KLqqqpCnYGAYYSp4QCNucfbWpvhcBRCQ0ZWWQjE=; b=CY5HDInZVU0qsTnbhn0TcewR0c/49onR73AZuVQLgAwDE1wpAB9It5sHyY2cmmOkRc Rb4s5Fj6xj+W3GwEoBAHg+33DwCk1RC9ZXt4/Xh76JvjtROKqIG5lVAau/NHIMYmC0cc 7qdGUkE3g+tOGHZnDOOowlEChCiufRRm2eY8XF5LA3xxu3SXzTJoVaS3/lrYtfqGfoCC ScKfdeu2nG/VLL3orp3/cBtVxSWpys1i7S4b3HEbdCItSWzbzOE4iuAF/fKhtV5JNpqH x5Dh1z14o1eckSRyTgquyhv0DcJVw6WYTEpWArSvsArQ/JNZ8ajfak7zmV+GqT7nToJ6 btvg== X-Forwarded-Encrypted: i=1; AJvYcCX9G94WDVAd+V1OCJoyqVulUTDLf+usSI08LRDMs3DgsBXTTAU/uLV3m6ePy85VV1WvHyZZV9QwP3zwXbQ=@vger.kernel.org X-Gm-Message-State: AOJu0YxlZLLBf2MBnfl2vkiw5APELchT7R/IEvuQ7T/bQrBL4dYJQ7bi SFp/5ssSP0zmBgaVdkmLKGYq1y/VuKZqwYrXXpZ7SFH7PLV09Qz83DDmLFt7Kyz2lAul3UMSQZb JO4ardS+3Aw== X-Google-Smtp-Source: AGHT+IEfB9gmIn13KKEoEvC6xYN1PwhauXtuIutekujJz/K9ZdS4TR7ttOYcI/LCLEtFVkiDzv/XXocZDyQw X-Received: from pjee16.prod.google.com ([2002:a17:90b:5790:b0:321:c567:44cf]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:903:22cc:b0:234:8a4a:adad with SMTP id d9443c01a7336-24478f71be3mr138270325ad.26.1755543926520; Mon, 18 Aug 2025 12:05:26 -0700 (PDT) Date: Mon, 18 Aug 2025 12:03:59 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-4-irogers@google.com> Subject: [PATCH v2 03/20] perf vendor events: Update arrowlake events/metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update events from v1.09 to v1.12. Update metrics from TMA 5.0 to 5.1. The event updates come from: https://github.com/intel/perfmon/commit/d5b1d2e9ee89035198ab7752e0378acba86= 907b6 https://github.com/intel/perfmon/commit/b9c162b7c98874321af36a411ee3c2194ea= 9afeb Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/arrowlake/arl-metrics.json | 180 +++++++++++------- .../pmu-events/arch/x86/arrowlake/cache.json | 122 ++++++++---- .../arch/x86/arrowlake/frontend.json | 40 ++-- .../pmu-events/arch/x86/arrowlake/memory.json | 22 +-- .../arch/x86/arrowlake/pipeline.json | 94 +++++---- tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 6 files changed, 282 insertions(+), 178 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/arrowlake/arl-metrics.json b/to= ols/perf/pmu-events/arch/x86/arrowlake/arl-metrics.json index b22a02450e6c..4f1f77404943 100644 --- a/tools/perf/pmu-events/arch/x86/arrowlake/arl-metrics.json +++ b/tools/perf/pmu-events/arch/x86/arrowlake/arl-metrics.json @@ -1,56 +1,56 @@ [ { "BriefDescription": "C10 residency percent per package", - "MetricExpr": "cstate_pkg@c10\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c10\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C10_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C8 residency percent per package", - "MetricExpr": "cstate_pkg@c8\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c8\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C8_Pkg_Residency", "ScaleUnit": "100%" @@ -567,7 +567,7 @@ }, { "BriefDescription": "Average CPU Utilization", - "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.REF_TSC@ / TSC", + "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.REF_TSC@ / msr@tsc\\,cpu= =3Dcpu_atom@", "MetricName": "tma_info_system_cpu_utilization", "Unit": "cpu_atom" }, @@ -774,7 +774,7 @@ { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-be\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-be\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BvOB;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", @@ -786,7 +786,7 @@ { "BriefDescription": "This category represents fraction of slots wa= sted due to incorrect speculations", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-bad\\-spec@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-bad\\-spec@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_bad_speculation", "MetricThreshold": "tma_bad_speculation > 0.15", @@ -812,36 +812,36 @@ "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)", "Unit": "cpu_core" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: ", + "Unit": "cpu_core" + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_capacity + tma_l1_latency_dependency + tma_lock_latency + tma_split_= loads + tma_store_fwd_blk)))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_capacity + tma_l1_latency_dependency + tma_lock_latency + tma_split_= loads + tma_store_early_blk + tma_store_fwd_blk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full", "Unit": "cpu_core" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_capacity= + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_sto= re_fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_= bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_latency_c= apacity / (tma_dtlb_load + tma_fb_full + tma_l1_latency_capacity + tma_l1_l= atency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)= ) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma= _l2_bound + tma_l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_dtl= b_load + tma_fb_full + tma_l1_latency_capacity + tma_l1_latency_dependency = + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bou= nd * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_split_loads / (tma_dtlb_load + tma_fb_ful= l + tma_l1_latency_capacity + tma_l1_latency_dependency + tma_lock_latency = + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bou= nd / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_sto= re_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tma_= split_stores + tma_store_latency + tma_streaming_stores)) + tma_memory_boun= d * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store + tma_f= alse_sharing + tma_split_stores + tma_store_latency + tma_streaming_stores)= ))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_capacity= + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_sto= re_early_blk + tma_store_fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tm= a_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound= )) * (tma_l1_latency_capacity / (tma_dtlb_load + tma_fb_full + tma_l1_laten= cy_capacity + tma_l1_latency_dependency + tma_lock_latency + tma_split_load= s + tma_store_early_blk + tma_store_fwd_blk)) + tma_memory_bound * (tma_l1_= bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_= store_bound)) * (tma_lock_latency / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_capacity + tma_l1_latency_dependency + tma_lock_latency + tma_split_= loads + tma_store_early_blk + tma_store_fwd_blk)) + tma_memory_bound * (tma= _l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + = tma_store_bound)) * (tma_split_loads / (tma_dtlb_load + tma_fb_full + tma_l= 1_latency_capacity + tma_l1_latency_dependency + tma_lock_latency + tma_spl= it_loads + tma_store_early_blk + tma_store_fwd_blk)) + tma_memory_bound * (= tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bo= und + tma_store_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_s= haring + tma_split_stores + tma_store_latency + tma_streaming_stores)) + tm= a_memory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2= _bound + tma_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_= store + tma_false_sharing + tma_split_stores + tma_store_latency + tma_stre= aming_stores)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency", "Unit": "cpu_core" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: ", - "Unit": "cpu_core" - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - c= pu_core@INST_RETIRED.REP_ITERATION@ / cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D= 1@) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cl= ears_resteers + tma_mispredicts_resteers * tma_other_mispredicts / tma_bran= ch_mispredicts) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unk= nown_branches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_miss= es + tma_itlb_misses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * t= ma_ms / (tma_dsb + tma_lsd + tma_mite + tma_ms))) - tma_bottleneck_big_code= ", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - c= pu_core@INST_RETIRED.REP_ITERATION@ / cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D= 1@) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cl= ears_resteers + tma_mispredicts_resteers * tma_other_mispredicts / tma_bran= ch_mispredicts) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unk= nown_branches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_miss= es + tma_itlb_misses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottlene= ck_big_code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20", @@ -849,7 +849,7 @@ }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * ((1 - cpu_core@INST_RETIRED.REP_ITERATION@ / = cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_restee= rs * tma_other_mispredicts / tma_branch_mispredicts) / (tma_clears_resteers= + tma_mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers= + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_m= s_switches) + tma_fetch_bandwidth * tma_ms / (tma_dsb + tma_lsd + tma_mite = + tma_ms)) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_bra= nch_mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_n= ukes / tma_other_nukes + tma_core_bound * (tma_serializing_operation + cpu_= core@RS.EMPTY_RESOURCE@ / tma_info_thread_clks * tma_ports_utilized_0) / (t= ma_divider + tma_ports_utilization + tma_serializing_operation) + tma_micro= code_sequencer / (tma_microcode_sequencer + tma_few_uops_instructions) * (t= ma_assists / tma_microcode_sequencer) * tma_heavy_operations)", + "MetricExpr": "100 * ((1 - cpu_core@INST_RETIRED.REP_ITERATION@ / = cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_restee= rs * tma_other_mispredicts / tma_branch_mispredicts) / (tma_clears_resteers= + tma_mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers= + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_m= s_switches) + tma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredict= s / tma_branch_mispredicts * tma_branch_mispredicts + tma_machine_clears * = tma_other_nukes / tma_other_nukes + tma_core_bound * (tma_serializing_opera= tion + cpu_core@RS.EMPTY_RESOURCE@ / tma_info_thread_clks * tma_ports_utili= zed_0) / (tma_divider + tma_ports_utilization + tma_serializing_operation) = + tma_microcode_sequencer / (tma_microcode_sequencer + tma_few_uops_instruc= tions) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -858,7 +858,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", - "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / (tma_dram= _bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (= tma_dtlb_load / (tma_dtlb_load + tma_fb_full + tma_l1_latency_capacity + tm= a_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fw= d_blk)) + tma_memory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bo= und + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_dtlb_store / (= tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_store_latency += tma_streaming_stores)))", + "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / (tma_dram= _bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (= tma_dtlb_load / (tma_dtlb_load + tma_fb_full + tma_l1_latency_capacity + tm= a_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_ea= rly_blk + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_= dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound))= * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_split_stores= + tma_store_latency + tma_streaming_stores)))", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", "MetricThreshold": "tma_bottleneck_memory_data_tlbs > 20", @@ -885,7 +885,7 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -902,7 +902,7 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU has wasted due to Branch Misprediction", - "MetricExpr": "cpu_core@topdown\\-br\\-mispredict@ / (cpu_core@top= down\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-re= tiring@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-br\\-mispredict@ / (cpu_core@top= down\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-re= tiring@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BadSpec;BrMispredicts;BvMP;TmaL2;TopdownL2;tma_L2_= group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_specula= tion > 0.15", @@ -1042,7 +1042,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to contested acces= ses", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS@ * min(= cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS@R, 24 * tma_info_system_core_fre= quency) + cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_HITM@ * min(cpu_core@MEM_LO= AD_L3_HIT_RETIRED.XSNP_HITM@R, 25 * tma_info_system_core_frequency)) * (1 += cpu_core@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2= ) / tma_info_thread_clks", "MetricGroup": "BvMS;DataSharing;LockCont;Offcore;Snoop;TopdownL4;= tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", @@ -1095,7 +1094,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to DSB (decoded uop cache) fetch pipe= line", - "MetricExpr": "(cpu_core@IDQ.DSB_UOPS\\,cmask\\=3D0x8\\,inv\\=3D0x= 1@ + cpu_core@IDQ.DSB_UOPS@ / (cpu_core@IDQ.DSB_UOPS@ + cpu_core@IDQ.MITE_U= OPS@) * (cpu_core@IDQ_BUBBLES.CYCLES_0_UOPS_DELIV.CORE@ - cpu_core@IDQ_BUBB= LES.FETCH_LATENCY@)) / tma_info_thread_clks", + "MetricExpr": "(cpu_core@IDQ.DSB_UOPS\\,cmask\\=3D0x8\\,inv\\=3D0x= 1@ / 2 + cpu_core@IDQ.DSB_UOPS@ / (cpu_core@IDQ.DSB_UOPS@ + cpu_core@IDQ.MI= TE_UOPS@) * (cpu_core@IDQ_BUBBLES.STARVATION_CYCLES@ - cpu_core@IDQ_BUBBLES= .FETCH_LATENCY@)) / tma_info_thread_clks", "MetricGroup": "DSB;FetchBW;TopdownL3;tma_L3_group;tma_fetch_bandw= idth_group", "MetricName": "tma_dsb", "MetricThreshold": "tma_dsb > 0.15 & tma_fetch_bandwidth > 0.2", @@ -1149,7 +1148,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -1166,7 +1165,7 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU was stalled due to Frontend latency issues", - "MetricExpr": "cpu_core@topdown\\-fetch\\-lat@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-fetch\\-lat@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Frontend;TmaL2;TopdownL2;tma_L2_group;tma_frontend= _bound_group", "MetricName": "tma_fetch_latency", "MetricThreshold": "tma_fetch_latency > 0.1 & tma_frontend_bound >= 0.15", @@ -1216,7 +1215,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-= point (FP) scalar uops fraction the CPU has retired", - "MetricExpr": "cpu_core@FP_ARITH_INST_RETIRED.SCALAR@ / (tma_retir= ing * tma_info_thread_slots)", + "MetricExpr": "cpu_core@FP_ARITH_OPS_RETIRED.SCALAR@ / (tma_retiri= ng * tma_info_thread_slots)", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_= group;tma_issue2P", "MetricName": "tma_fp_scalar", "MetricThreshold": "tma_fp_scalar > 0.1 & (tma_fp_arith > 0.2 & tm= a_light_operations > 0.6)", @@ -1226,7 +1225,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-= point (FP) vector uops fraction the CPU has retired aggregated across all v= ector widths", - "MetricExpr": "cpu_core@FP_ARITH_INST_RETIRED.VECTOR@ / (tma_retir= ing * tma_info_thread_slots)", + "MetricExpr": "cpu_core@FP_ARITH_OPS_RETIRED.VECTOR@ / (tma_retiri= ng * tma_info_thread_slots)", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_= group;tma_issue2P", "MetricName": "tma_fp_vector", "MetricThreshold": "tma_fp_vector > 0.1 & (tma_fp_arith > 0.2 & tm= a_light_operations > 0.6)", @@ -1236,7 +1235,7 @@ }, { "BriefDescription": "This metric approximates arithmetic FP vector= uops fraction the CPU has retired for 128-bit wide vectors", - "MetricExpr": "(cpu_core@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE@= + cpu_core@FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE@) / (tma_retiring * tm= a_info_thread_slots)", + "MetricExpr": "(cpu_core@FP_ARITH_OPS_RETIRED.128B_PACKED_DOUBLE@ = + cpu_core@FP_ARITH_OPS_RETIRED.128B_PACKED_SINGLE@) / (tma_retiring * tma_= info_thread_slots)", "MetricGroup": "Compute;Flops;TopdownL5;tma_L5_group;tma_fp_vector= _group;tma_issue2P", "MetricName": "tma_fp_vector_128b", "MetricThreshold": "tma_fp_vector_128b > 0.1 & (tma_fp_vector > 0.= 1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6))", @@ -1246,7 +1245,7 @@ }, { "BriefDescription": "This metric approximates arithmetic FP vector= uops fraction the CPU has retired for 256-bit wide vectors", - "MetricExpr": "cpu_core@FP_ARITH_INST_RETIRED.VECTOR\\,umask\\=3D0= x30@ / (tma_retiring * tma_info_thread_slots)", + "MetricExpr": "cpu_core@FP_ARITH_OPS_RETIRED.VECTOR\\,umask\\=3D0x= 30@ / (tma_retiring * tma_info_thread_slots)", "MetricGroup": "Compute;Flops;TopdownL5;tma_L5_group;tma_fp_vector= _group;tma_issue2P", "MetricName": "tma_fp_vector_256b", "MetricThreshold": "tma_fp_vector_256b > 0.1 & (tma_fp_vector > 0.= 1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6))", @@ -1257,7 +1256,7 @@ { "BriefDescription": "This category represents fraction of slots wh= ere the processor's Frontend undersupplies its Backend", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-fe\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-fe\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BvFB;BvIO;Default;PGO;TmaL1;TopdownL1;tma_L1_group= ", "MetricName": "tma_frontend_bound", "MetricThreshold": "tma_frontend_bound > 0.15", @@ -1278,7 +1277,7 @@ }, { "BriefDescription": "This metric represents fraction of slots wher= e the CPU was retiring heavy-weight operations -- instructions that require= two or more uops or micro-coded sequences", - "MetricExpr": "cpu_core@topdown\\-heavy\\-ops@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-heavy\\-ops@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Retire;TmaL2;TopdownL2;tma_L2_group;tma_retiring_g= roup", "MetricName": "tma_heavy_operations", "MetricThreshold": "tma_heavy_operations > 0.1", @@ -1456,7 +1455,7 @@ }, { "BriefDescription": "Floating Point Operations Per Cycle", - "MetricExpr": "(cpu_core@FP_ARITH_INST_RETIRED.SCALAR@ + 2 * cpu_c= ore@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE@ + 4 * cpu_core@FP_ARITH_INST_= RETIRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE@) = / tma_info_thread_clks", + "MetricExpr": "(cpu_core@FP_ARITH_OPS_RETIRED.SCALAR@ + 2 * cpu_co= re@FP_ARITH_OPS_RETIRED.128B_PACKED_DOUBLE@ + 4 * cpu_core@FP_ARITH_OPS_RET= IRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_OPS_RETIRED.256B_PACKED_SINGLE@) / tm= a_info_thread_clks", "MetricGroup": "Flops;Ret", "MetricName": "tma_info_core_flopc", "Unit": "cpu_core" @@ -1597,7 +1596,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic instruction (l= ower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_INS= T_RETIRED.SCALAR@ + cpu_core@FP_ARITH_INST_RETIRED.VECTOR@)", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_OPS= _RETIRED.SCALAR@ + cpu_core@FP_ARITH_OPS_RETIRED.VECTOR@)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_iparith", "MetricThreshold": "tma_info_inst_mix_iparith < 10", @@ -1606,7 +1605,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic AVX/SSE 128-bi= t instruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_INS= T_RETIRED.128B_PACKED_DOUBLE@ + cpu_core@FP_ARITH_INST_RETIRED.128B_PACKED_= SINGLE@)", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_OPS= _RETIRED.128B_PACKED_DOUBLE@ + cpu_core@FP_ARITH_OPS_RETIRED.128B_PACKED_SI= NGLE@)", "MetricGroup": "Flops;FpVector;InsType", "MetricName": "tma_info_inst_mix_iparith_avx128", "MetricThreshold": "tma_info_inst_mix_iparith_avx128 < 10", @@ -1615,7 +1614,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic AVX* 256-bit i= nstruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_INS= T_RETIRED.256B_PACKED_DOUBLE@ + cpu_core@FP_ARITH_INST_RETIRED.256B_PACKED_= SINGLE@)", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_OPS= _RETIRED.256B_PACKED_DOUBLE@ + cpu_core@FP_ARITH_OPS_RETIRED.256B_PACKED_SI= NGLE@)", "MetricGroup": "Flops;FpVector;InsType", "MetricName": "tma_info_inst_mix_iparith_avx256", "MetricThreshold": "tma_info_inst_mix_iparith_avx256 < 10", @@ -1624,7 +1623,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic Scalar Double-= Precision instruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FP_ARITH_INST= _RETIRED.SCALAR_DOUBLE@", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FP_ARITH_OPS_= RETIRED.SCALAR_DOUBLE@", "MetricGroup": "Flops;FpScalar;InsType", "MetricName": "tma_info_inst_mix_iparith_scalar_dp", "MetricThreshold": "tma_info_inst_mix_iparith_scalar_dp < 10", @@ -1633,7 +1632,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic Scalar Single-= Precision instruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FP_ARITH_INST= _RETIRED.SCALAR_SINGLE@", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FP_ARITH_OPS_= RETIRED.SCALAR_SINGLE@", "MetricGroup": "Flops;FpScalar;InsType", "MetricName": "tma_info_inst_mix_iparith_scalar_sp", "MetricThreshold": "tma_info_inst_mix_iparith_scalar_sp < 10", @@ -1658,7 +1657,7 @@ }, { "BriefDescription": "Instructions per Floating Point (FP) Operatio= n (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_INS= T_RETIRED.SCALAR@ + 2 * cpu_core@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE@ = + 4 * cpu_core@FP_ARITH_INST_RETIRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_INST_= RETIRED.256B_PACKED_SINGLE@)", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_OPS= _RETIRED.SCALAR@ + 2 * cpu_core@FP_ARITH_OPS_RETIRED.128B_PACKED_DOUBLE@ + = 4 * cpu_core@FP_ARITH_OPS_RETIRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_OPS_RETI= RED.256B_PACKED_SINGLE@)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_ipflop", "MetricThreshold": "tma_info_inst_mix_ipflop < 10", @@ -1713,7 +1712,7 @@ }, { "BriefDescription": "Average per-thread data fill bandwidth to the= L1 data cache [GB / sec]", - "MetricExpr": "64 * cpu_core@L1D.REPLACEMENT@ / 1e9 / tma_info_sys= tem_time", + "MetricExpr": "64 * cpu_core@L1D.L1_REPLACEMENT@ / 1e9 / tma_info_= system_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l1d_cache_fill_bw", "Unit": "cpu_core" @@ -1725,6 +1724,13 @@ "MetricName": "tma_info_memory_l1dl0_cache_fill_bw", "Unit": "cpu_core" }, + { + "BriefDescription": "L0 cache true misses per kilo instruction for= retired demand loads", + "MetricExpr": "1e3 * (cpu_core@MEM_LOAD_RETIRED.L1_MISS@ + cpu_cor= e@MEM_LOAD_RETIRED.L1_HIT_L1@) / cpu_core@INST_RETIRED.ANY@", + "MetricGroup": "CacheHits;Mem", + "MetricName": "tma_info_memory_l1dl0_mpki", + "Unit": "cpu_core" + }, { "BriefDescription": "L1 cache true misses per kilo instruction for= retired demand loads", "MetricExpr": "1e3 * cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / cpu_core= @INST_RETIRED.ANY@", @@ -1940,6 +1946,13 @@ "MetricName": "tma_info_pipeline_fetch_mite", "Unit": "cpu_core" }, + { + "BriefDescription": "Average number of uops fetched from MS per cy= cle", + "MetricExpr": "cpu_core@IDQ.MS_UOPS@ / cpu_core@IDQ.MS_UOPS\\,cmas= k\\=3D1@", + "MetricGroup": "Fed;FetchLat;MicroSeq", + "MetricName": "tma_info_pipeline_fetch_ms", + "Unit": "cpu_core" + }, { "BriefDescription": "Instructions per a microcode Assist invocatio= n", "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@ASSISTS.ANY@", @@ -1974,7 +1987,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc\\,cpu= =3Dcpu_core@ / 1e9 / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency", "Unit": "cpu_core" @@ -1988,14 +2001,22 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.REF_TSC@ / TSC", + "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.REF_TSC@ / msr@tsc\\,cpu= =3Dcpu_core@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized", "Unit": "cpu_core" }, + { + "BriefDescription": "Average external Memory Bandwidth Use for rea= ds and writes [GB / sec]", + "MetricExpr": "32 * UNC_M_TOTAL_DATA / 1e9 / tma_info_system_time", + "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", + "MetricName": "tma_info_system_dram_bw_use", + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full", + "Unit": "cpu_core" + }, { "BriefDescription": "Giga Floating Point Operations Per Second", - "MetricExpr": "(cpu_core@FP_ARITH_INST_RETIRED.SCALAR@ + 2 * cpu_c= ore@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE@ + 4 * cpu_core@FP_ARITH_INST_= RETIRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE@) = / 1e9 / tma_info_system_time", + "MetricExpr": "(cpu_core@FP_ARITH_OPS_RETIRED.SCALAR@ + 2 * cpu_co= re@FP_ARITH_OPS_RETIRED.128B_PACKED_DOUBLE@ + 4 * cpu_core@FP_ARITH_OPS_RET= IRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_OPS_RETIRED.256B_PACKED_SINGLE@) / 1e= 9 / tma_info_system_time", "MetricGroup": "Cor;Flops;HPC", "MetricName": "tma_info_system_gflops", "PublicDescription": "Giga Floating Point Operations Per Second. A= ggregate across all supported options of: FP precisions, scalar and vector = instructions, vector-width", @@ -2061,6 +2082,13 @@ "MetricName": "tma_info_system_turbo_utilization", "Unit": "cpu_core" }, + { + "BriefDescription": "Measured Average Uncore Frequency for the SoC= [GHz]", + "MetricExpr": "tma_info_system_socket_clks / 1e9 / tma_info_system= _time", + "MetricGroup": "SoC", + "MetricName": "tma_info_system_uncore_frequency", + "Unit": "cpu_core" + }, { "BriefDescription": "Per-Logical Processor actual clocks when the = Logical Processor is active.", "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.THREAD@", @@ -2182,12 +2210,12 @@ "Unit": "cpu_core" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", - "MetricExpr": "4 * cpu_core@DEPENDENT_LOADS.ANY@ / tma_info_thread= _clks", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", + "MetricExpr": "4 * cpu_core@DEPENDENT_LOADS.ANY\\,cmask\\=3D1@ / t= ma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2203,7 +2231,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles wit= h demand load accesses that hit the L2 cache under unloaded scenarios (poss= ibly L2 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu_core@MEM_LOAD_RETIRED.L2_HIT@ * min(cpu_core@ME= M_LOAD_RETIRED.L2_HIT@R, 3 * tma_info_system_core_frequency) * (1 + cpu_cor= e@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2) / tma_= info_thread_clks", "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_l2_bound_grou= p", "MetricName": "tma_l2_hit_latency", @@ -2224,12 +2251,11 @@ }, { "BriefDescription": "This metric estimates fraction of cycles with= demand load accesses that hit the L3 cache under unloaded scenarios (possi= bly L3 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu_core@MEM_LOAD_RETIRED.L3_HIT@ * min(cpu_core@ME= M_LOAD_RETIRED.L3_HIT@R, 9 * tma_info_system_core_frequency) * (1 + cpu_cor= e@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2) / tma_= info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2311,6 +2337,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU spent handling cache misses due to lock operations", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu_core@MEM_INST_RETIRED.LOCK_LOADS@ * cpu_core@ME= M_INST_RETIRED.LOCK_LOADS@R / tma_info_thread_clks", "MetricGroup": "LockCont;Offcore;TopdownL4;tma_L4_group;tma_issueR= FO;tma_l1_bound_group", "MetricName": "tma_lock_latency", @@ -2321,7 +2348,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to LSD (Loop Stream Detector) unit", - "MetricExpr": "cpu_core@LSD.UOPS\\,cmask\\=3D0x8\\,inv\\=3D0x1@ / = tma_info_thread_clks", + "MetricExpr": "cpu_core@LSD.UOPS\\,cmask\\=3D0x8\\,inv\\=3D0x1@ / = tma_info_thread_clks / 2", "MetricGroup": "FetchBW;LSD;TopdownL3;tma_L3_group;tma_fetch_bandw= idth_group", "MetricName": "tma_lsd", "MetricThreshold": "tma_lsd > 0.15 & tma_fetch_bandwidth > 0.2", @@ -2346,7 +2373,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2356,13 +2383,13 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%", "Unit": "cpu_core" }, { "BriefDescription": "This metric represents fraction of slots the = Memory subsystem within the Backend was a bottleneck", - "MetricExpr": "cpu_core@topdown\\-mem\\-bound@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-mem\\-bound@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Backend;TmaL2;TopdownL2;tma_L2_group;tma_backend_b= ound_group", "MetricName": "tma_memory_bound", "MetricThreshold": "tma_memory_bound > 0.2 & tma_backend_bound > 0= .2", @@ -2373,7 +2400,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to LFENCE Instructions.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * cpu_core@MISC2_RETIRED.LFENCE@ / tma_info_thre= ad_clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_memory_fence", @@ -2412,7 +2438,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the MITE pipeline (the legacy deco= de pipeline)", - "MetricExpr": "(cpu_core@IDQ.MITE_UOPS\\,cmask\\=3D0x8\\,inv\\=3D0= x1@ / 2 + cpu_core@IDQ.MITE_UOPS@ / (cpu_core@IDQ.DSB_UOPS@ + cpu_core@IDQ.= MITE_UOPS@) * (cpu_core@IDQ_BUBBLES.CYCLES_0_UOPS_DELIV.CORE@ - cpu_core@ID= Q_BUBBLES.FETCH_LATENCY@)) / tma_info_thread_clks", + "MetricExpr": "(cpu_core@IDQ.MITE_UOPS\\,cmask\\=3D0x8\\,inv\\=3D0= x1@ / 2 + cpu_core@IDQ.MITE_UOPS@ / (cpu_core@IDQ.DSB_UOPS@ + cpu_core@IDQ.= MITE_UOPS@) * (cpu_core@IDQ_BUBBLES.STARVATION_CYCLES@ - cpu_core@IDQ_BUBBL= ES.FETCH_LATENCY@)) / tma_info_thread_clks", "MetricGroup": "DSBmiss;FetchBW;TopdownL3;tma_L3_group;tma_fetch_b= andwidth_group", "MetricName": "tma_mite", "MetricThreshold": "tma_mite > 0.1 & tma_fetch_bandwidth > 0.2", @@ -2432,7 +2458,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "cpu_core@IDQ.MS_CYCLES_ANY@ / tma_info_thread_clks", + "MetricExpr": "cpu_core@IDQ.MS_CYCLES_ANY@ / tma_info_thread_clks = / 1.8", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -2471,7 +2497,8 @@ }, { "BriefDescription": "This metric represents the remaining light uo= ps fraction the CPU has executed - remaining means not covered by other sib= ling nodes", - "MetricExpr": "max(0, tma_light_operations - (tma_x87_use + (cpu_c= ore@FP_ARITH_INST_RETIRED.SCALAR@ + cpu_core@FP_ARITH_INST_RETIRED.VECTOR@)= / (tma_retiring * tma_info_thread_slots) + (cpu_core@INT_VEC_RETIRED.ADD_1= 28@ + cpu_core@INT_VEC_RETIRED.VNNI_128@ + cpu_core@INT_VEC_RETIRED.ADD_256= @ + cpu_core@INT_VEC_RETIRED.MUL_256@ + cpu_core@INT_VEC_RETIRED.VNNI_256@)= / (tma_retiring * tma_info_thread_slots) + tma_memory_operations + tma_fus= ed_instructions + tma_non_fused_branches))", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "max(0, tma_light_operations - (tma_x87_use + (cpu_c= ore@FP_ARITH_OPS_RETIRED.SCALAR@ + cpu_core@FP_ARITH_OPS_RETIRED.VECTOR@) /= (tma_retiring * tma_info_thread_slots) + (cpu_core@INT_VEC_RETIRED.ADD_128= @ + cpu_core@INT_VEC_RETIRED.VNNI_128@ + cpu_core@INT_VEC_RETIRED.ADD_256@ = + cpu_core@INT_VEC_RETIRED.MUL_256@ + cpu_core@INT_VEC_RETIRED.VNNI_256@) /= (tma_retiring * tma_info_thread_slots) + tma_memory_operations + tma_fused= _instructions + tma_non_fused_branches))", "MetricGroup": "Pipeline;TopdownL3;tma_L3_group;tma_light_operatio= ns_group", "MetricName": "tma_other_light_ops", "MetricThreshold": "tma_other_light_ops > 0.3 & tma_light_operatio= ns > 0.6", @@ -2509,6 +2536,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "((cpu_core@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ + (cpu_c= ore@EXE_ACTIVITY.1_PORTS_UTIL@ + tma_retiring * cpu_core@EXE_ACTIVITY.2_3_P= ORTS_UTIL@)) / tma_info_thread_clks if cpu_core@ARITH.DIV_ACTIVE@ < cpu_cor= e@CYCLE_ACTIVITY.STALLS_TOTAL@ - cpu_core@EXE_ACTIVITY.BOUND_ON_LOADS@ else= (cpu_core@EXE_ACTIVITY.1_PORTS_UTIL@ + tma_retiring * cpu_core@EXE_ACTIVIT= Y.2_3_PORTS_UTIL@) / tma_info_thread_clks)", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", @@ -2519,6 +2547,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "cpu_core@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ / tma_info= _thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", @@ -2529,6 +2558,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "cpu_core@EXE_ACTIVITY.1_PORTS_UTIL@ / tma_info_thre= ad_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", @@ -2539,7 +2569,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@EXE_ACTIVITY.2_PORTS_UTIL@ / tma_info_thre= ad_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", @@ -2550,7 +2579,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 3 or more uops per cycle on all execution ports (Logical= Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@UOPS_EXECUTED.CYCLES_GE_3@ / tma_info_thre= ad_clks", "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_ut= ilization_group", "MetricName": "tma_ports_utilized_3m", @@ -2571,7 +2599,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-retiring@ / (cpu_core@topdown\\-= fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@= + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-retiring@ / (cpu_core@topdown\\-= fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@= + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -2586,7 +2614,7 @@ "MetricGroup": "BvIO;PortsUtil;TopdownL3;tma_L3_group;tma_core_bou= nd_group;tma_issueSO", "MetricName": "tma_serializing_operation", "MetricThreshold": "tma_serializing_operation > 0.1 & (tma_core_bo= und > 0.1 & tma_backend_bound > 0.2)", - "PublicDescription": "This metric represents fraction of cycles th= e CPU issue-pipeline was stalled due to serializing operations. Instruction= s like CPUID; WRMSR or LFENCE serialize the out-of-order execution which ma= y limit performance. Sample with: RESOURCE_STALLS.SCOREBOARD. Related metri= cs: tma_ms_switches", + "PublicDescription": "This metric represents fraction of cycles th= e CPU issue-pipeline was stalled due to serializing operations. Instruction= s like CPUID; WRMSR or LFENCE serialize the out-of-order execution which ma= y limit performance. Sample with: PARTIAL_RAT_STALLS.SCOREBOARD. Related me= trics: tma_ms_switches", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2602,7 +2630,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to PAUSE Instructions", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.PAUSE@ / tma_info_thread_= clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_slow_pause", @@ -2637,7 +2664,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2651,6 +2678,15 @@ "ScaleUnit": "100%", "Unit": "cpu_core" }, + { + "BriefDescription": "This metric estimates clocks wasted due to lo= ads blocked due to unknown store address (did not do memory disambiguation)= or due to unknown store data", + "MetricExpr": "7 * cpu_core@LD_BLOCKS.STORE_EARLY\\,cmask\\=3D1@ /= tma_info_thread_clks", + "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", + "MetricName": "tma_store_early_blk", + "MetricThreshold": "tma_store_early_blk > 0.2", + "ScaleUnit": "100%", + "Unit": "cpu_core" + }, { "BriefDescription": "This metric roughly estimates fraction of cyc= les when the memory subsystem had loads blocked since they could not forwar= d data from earlier (in program order) overlapping stores", "MetricExpr": "13 * cpu_core@LD_BLOCKS.STORE_FORWARD@ / tma_info_t= hread_clks", diff --git a/tools/perf/pmu-events/arch/x86/arrowlake/cache.json b/tools/pe= rf/pmu-events/arch/x86/arrowlake/cache.json index 91929d8bcf47..f5168b55a6f4 100644 --- a/tools/perf/pmu-events/arch/x86/arrowlake/cache.json +++ b/tools/perf/pmu-events/arch/x86/arrowlake/cache.json @@ -28,6 +28,16 @@ "UMask": "0x1", "Unit": "cpu_core" }, + { + "BriefDescription": "Cachelines replaced into the L1 d-cache. Succ= essful replacements only (not blocked) and exclude WB-miss case", + "Counter": "0,1,2,3,4,5,6,7,8,9", + "EventCode": "0x51", + "EventName": "L1D.L1_REPLACEMENT", + "PublicDescription": "Counts cachelines replaced into the L1 d-cac= he.", + "SampleAfterValue": "1000003", + "UMask": "0x4", + "Unit": "cpu_core" + }, { "BriefDescription": "Cachelines replaced into the L0 and L1 d-cach= e. Successful replacements only (not blocked) and exclude WB-miss case", "Counter": "0,1,2,3,4,5,6,7,8,9", @@ -540,7 +550,7 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.ALL_LOADS", - "PublicDescription": "Counts Instructions with at least one archit= ecturally visible load retired. Available PDIST counters: 0", + "PublicDescription": "Counts Instructions with at least one archit= ecturally visible load retired. Available PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x81", "Unit": "cpu_core" @@ -551,7 +561,7 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.ALL_STORES", - "PublicDescription": "Counts all retired store instructions. Avail= able PDIST counters: 0", + "PublicDescription": "Counts all retired store instructions. Avail= able PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x82", "Unit": "cpu_core" @@ -561,7 +571,7 @@ "Counter": "0,1,2,3", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.ALL_SWPF", - "PublicDescription": "Counts all retired software prefetch instruc= tions. Available PDIST counters: 0", + "PublicDescription": "Counts all retired software prefetch instruc= tions. Available PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x84", "Unit": "cpu_core" @@ -572,7 +582,7 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.ANY", - "PublicDescription": "Counts all retired memory instructions - loa= ds and stores. Available PDIST counters: 0", + "PublicDescription": "Counts all retired memory instructions - loa= ds and stores. Available PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x87", "Unit": "cpu_core" @@ -583,7 +593,7 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.LOCK_LOADS", - "PublicDescription": "Counts retired load instructions with locked= access. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with locked= access. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x21", "Unit": "cpu_core" @@ -594,7 +604,7 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.SPLIT_LOADS", - "PublicDescription": "Counts retired load instructions that split = across a cacheline boundary. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions that split = across a cacheline boundary. Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x41", "Unit": "cpu_core" @@ -605,18 +615,29 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.SPLIT_STORES", - "PublicDescription": "Counts retired store instructions that split= across a cacheline boundary. Available PDIST counters: 0", + "PublicDescription": "Counts retired store instructions that split= across a cacheline boundary. Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x42", "Unit": "cpu_core" }, + { + "BriefDescription": "Retired instructions that hit the STLB.", + "Counter": "0,1,2,3", + "Data_LA": "1", + "EventCode": "0xd0", + "EventName": "MEM_INST_RETIRED.STLB_HIT_ANY", + "PublicDescription": "Number of retired instructions with a clean = hit in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", + "SampleAfterValue": "100003", + "UMask": "0xf", + "Unit": "cpu_core" + }, { "BriefDescription": "Retired load instructions that hit the STLB.", "Counter": "0,1,2,3", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.STLB_HIT_LOADS", - "PublicDescription": "Number of retired load instructions with a c= lean hit in the 2nd-level TLB (STLB). Available PDIST counters: 0", + "PublicDescription": "Number of retired load instructions with a c= lean hit in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x9", "Unit": "cpu_core" @@ -627,18 +648,39 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.STLB_HIT_STORES", - "PublicDescription": "Number of retired store instructions that hi= t in the 2nd-level TLB (STLB). Available PDIST counters: 0", + "PublicDescription": "Number of retired store instructions that hi= t in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0xa", "Unit": "cpu_core" }, + { + "BriefDescription": "Retired SWPF instructions that hit the STLB.", + "Counter": "0,1,2,3", + "EventCode": "0xd0", + "EventName": "MEM_INST_RETIRED.STLB_HIT_SWPF", + "PublicDescription": "Number of retired SWPF instructions that hit= in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", + "SampleAfterValue": "1000003", + "UMask": "0xc", + "Unit": "cpu_core" + }, + { + "BriefDescription": "Retired instructions that miss the STLB.", + "Counter": "0,1,2,3", + "Data_LA": "1", + "EventCode": "0xd0", + "EventName": "MEM_INST_RETIRED.STLB_MISS_ANY", + "PublicDescription": "Retired instructions that miss the STLB. Ava= ilable PDIST counters: 0,1", + "SampleAfterValue": "100003", + "UMask": "0x17", + "Unit": "cpu_core" + }, { "BriefDescription": "Retired load instructions that miss the STLB.= ", "Counter": "0,1,2,3", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.STLB_MISS_LOADS", - "PublicDescription": "Number of retired load instructions that (st= art a) miss in the 2nd-level TLB (STLB). Available PDIST counters: 0", + "PublicDescription": "Number of retired load instructions that (st= art a) miss in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x11", "Unit": "cpu_core" @@ -649,18 +691,28 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.STLB_MISS_STORES", - "PublicDescription": "Number of retired store instructions that (s= tart a) miss in the 2nd-level TLB (STLB). Available PDIST counters: 0", + "PublicDescription": "Number of retired store instructions that (s= tart a) miss in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x12", "Unit": "cpu_core" }, + { + "BriefDescription": "Retired SWPF instructions that miss the STLB.= ", + "Counter": "0,1,2,3", + "EventCode": "0xd0", + "EventName": "MEM_INST_RETIRED.STLB_MISS_SWPF", + "PublicDescription": "Number of retired SWPF instructions that (st= art a) miss in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", + "SampleAfterValue": "1000003", + "UMask": "0x14", + "Unit": "cpu_core" + }, { "BriefDescription": "Retired load instructions whose data sources = were a cross-core Snoop hits and forwards data from an in on-package core c= ache (induced by NI$)", "Counter": "0,1,2,3", "Data_LA": "1", "EventCode": "0xd2", "EventName": "MEM_LOAD_L3_HIT_RETIRED.XSNP_FWD", - "PublicDescription": "Counts retired load instructions whose data = sources were a cross-core Snoop hits and forwards data from an in on-packag= e core cache (induced by NI$) Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions whose data = sources were a cross-core Snoop hits and forwards data from an in on-packag= e core cache (induced by NI$) Available PDIST counters: 0,1", "SampleAfterValue": "20011", "UMask": "0x10", "Unit": "cpu_core" @@ -671,7 +723,7 @@ "Data_LA": "1", "EventCode": "0xd2", "EventName": "MEM_LOAD_L3_HIT_RETIRED.XSNP_HITM", - "PublicDescription": "Counts retired load instructions whose data = sources were HitM responses from shared L3, Hit-with-FWD is normally exclud= ed. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions whose data = sources were HitM responses from shared L3, Hit-with-FWD is normally exclud= ed. Available PDIST counters: 0,1", "SampleAfterValue": "20011", "UMask": "0x4", "Unit": "cpu_core" @@ -682,7 +734,7 @@ "Data_LA": "1", "EventCode": "0xd2", "EventName": "MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS", - "PublicDescription": "Counts the retired load instructions whose d= ata sources were L3 hit and cross-core snoop missed in on-pkg core cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts the retired load instructions whose d= ata sources were L3 hit and cross-core snoop missed in on-pkg core cache. A= vailable PDIST counters: 0,1", "SampleAfterValue": "20011", "UMask": "0x1", "Unit": "cpu_core" @@ -693,7 +745,7 @@ "Data_LA": "1", "EventCode": "0xd2", "EventName": "MEM_LOAD_L3_HIT_RETIRED.XSNP_NO_FWD", - "PublicDescription": "Counts retired load instructions whose data = sources were L3 and cross-core snoop hits in on-pkg core cache. Available P= DIST counters: 0", + "PublicDescription": "Counts retired load instructions whose data = sources were L3 and cross-core snoop hits in on-pkg core cache. Available P= DIST counters: 0,1", "SampleAfterValue": "20011", "UMask": "0x2", "Unit": "cpu_core" @@ -704,7 +756,7 @@ "Data_LA": "1", "EventCode": "0xd4", "EventName": "MEM_LOAD_MISC_RETIRED.UC", - "PublicDescription": "Retired instructions with at least one load = to uncacheable memory-type, or at least one cache-line split locked access = (Bus Lock). Available PDIST counters: 0", + "PublicDescription": "Retired instructions with at least one load = to uncacheable memory-type, or at least one cache-line split locked access = (Bus Lock). Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x4", "Unit": "cpu_core" @@ -715,7 +767,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.FB_HIT", - "PublicDescription": "Counts retired load instructions with at lea= st one uop was load missed in L1 but hit FB (Fill Buffers) due to preceding= miss to the same cache line with data not ready. Available PDIST counters:= 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop was load missed in L1 but hit FB (Fill Buffers) due to preceding= miss to the same cache line with data not ready. Available PDIST counters:= 0,1", "SampleAfterValue": "100007", "UMask": "0x40", "Unit": "cpu_core" @@ -726,7 +778,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L1_HIT", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the L1 data cache. This event includes all SW prefet= ches and lock instructions regardless of the data source. Available PDIST c= ounters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the L1 data cache. This event includes all SW prefet= ches and lock instructions regardless of the data source. Available PDIST c= ounters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x101", "Unit": "cpu_core" @@ -737,7 +789,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L1_HIT_L0", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the Level 0 of the L1 data cache. This event include= s all SW prefetches and lock instructions regardless of the data source. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the Level 0 of the L1 data cache. This event include= s all SW prefetches and lock instructions regardless of the data source. Av= ailable PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -747,7 +799,7 @@ "Counter": "0,1,2,3", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L1_HIT_L1", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the Level 1 of the L1 data cache. Available PDIST co= unters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the Level 1 of the L1 data cache. Available PDIST co= unters: 0,1", "SampleAfterValue": "1000003", "Unit": "cpu_core" }, @@ -757,7 +809,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L1_MISS", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that missed in the L1 cache. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that missed in the L1 cache. Available PDIST counters: 0,1", "SampleAfterValue": "200003", "UMask": "0x8", "Unit": "cpu_core" @@ -768,7 +820,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L2_HIT", - "PublicDescription": "Counts retired load instructions with L2 cac= he hits as data sources. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with L2 cac= he hits as data sources. Available PDIST counters: 0,1", "SampleAfterValue": "200003", "UMask": "0x2", "Unit": "cpu_core" @@ -779,7 +831,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L2_MISS", - "PublicDescription": "Counts retired load instructions missed L2 c= ache as data sources. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions missed L2 c= ache as data sources. Available PDIST counters: 0,1", "SampleAfterValue": "100021", "UMask": "0x10", "Unit": "cpu_core" @@ -790,7 +842,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L3_HIT", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the L3 cache. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the L3 cache. Available PDIST counters: 0,1", "SampleAfterValue": "100021", "UMask": "0x4", "Unit": "cpu_core" @@ -801,7 +853,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L3_MISS", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that missed in the L3 cache. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that missed in the L3 cache. Available PDIST counters: 0,1", "SampleAfterValue": "50021", "UMask": "0x20", "Unit": "cpu_core" @@ -1029,7 +1081,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_1024", @@ -1053,7 +1105,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_128", @@ -1077,7 +1129,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_16", @@ -1089,7 +1141,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_2048", @@ -1113,7 +1165,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_256", @@ -1137,7 +1189,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_32", @@ -1161,7 +1213,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_4", @@ -1185,7 +1237,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_512", @@ -1209,7 +1261,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_64", @@ -1233,7 +1285,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_8", diff --git a/tools/perf/pmu-events/arch/x86/arrowlake/frontend.json b/tools= /perf/pmu-events/arch/x86/arrowlake/frontend.json index 56cf1ec63200..db2ef84ca041 100644 --- a/tools/perf/pmu-events/arch/x86/arrowlake/frontend.json +++ b/tools/perf/pmu-events/arch/x86/arrowlake/frontend.json @@ -81,7 +81,7 @@ "EventName": "FRONTEND_RETIRED.ANY_ANT", "MSRIndex": "0x3F7", "MSRValue": "0x9", - "PublicDescription": "Always Not Taken (ANT) conditional retired b= ranches (no BTB entry and not mispredicted) Available PDIST counters: 0", + "PublicDescription": "Always Not Taken (ANT) conditional retired b= ranches (no BTB entry and not mispredicted) Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -93,7 +93,7 @@ "EventName": "FRONTEND_RETIRED.ANY_DSB_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x1", - "PublicDescription": "Counts retired Instructions that experienced= DSB (Decode stream buffer i.e. the decoded instruction-cache) miss. Availa= ble PDIST counters: 0", + "PublicDescription": "Counts retired Instructions that experienced= DSB (Decode stream buffer i.e. the decoded instruction-cache) miss. Availa= ble PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -179,7 +179,7 @@ "EventName": "FRONTEND_RETIRED.DSB_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x11", - "PublicDescription": "Number of retired Instructions that experien= ced a critical DSB (Decode stream buffer i.e. the decoded instruction-cache= ) miss. Critical means stalls were exposed to the back-end as a result of t= he DSB miss. Available PDIST counters: 0", + "PublicDescription": "Number of retired Instructions that experien= ced a critical DSB (Decode stream buffer i.e. the decoded instruction-cache= ) miss. Critical means stalls were exposed to the back-end as a result of t= he DSB miss. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -218,7 +218,7 @@ "EventName": "FRONTEND_RETIRED.ITLB_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x14", - "PublicDescription": "Counts retired Instructions that experienced= iTLB (Instruction TLB) true miss. Available PDIST counters: 0", + "PublicDescription": "Counts retired Instructions that experienced= iTLB (Instruction TLB) true miss. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -239,7 +239,7 @@ "EventName": "FRONTEND_RETIRED.L1I_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x12", - "PublicDescription": "Counts retired Instructions who experienced = Instruction L1 Cache true miss. Available PDIST counters: 0", + "PublicDescription": "Counts retired Instructions who experienced = Instruction L1 Cache true miss. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -251,7 +251,7 @@ "EventName": "FRONTEND_RETIRED.L2_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x13", - "PublicDescription": "Counts retired Instructions who experienced = Instruction L2 Cache true miss. Available PDIST counters: 0", + "PublicDescription": "Counts retired Instructions who experienced = Instruction L2 Cache true miss. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -263,7 +263,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_128", "MSRIndex": "0x3F7", "MSRValue": "0x608006", - "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 12= 8 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0", + "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 12= 8 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -275,7 +275,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_16", "MSRIndex": "0x3F7", "MSRValue": "0x601006", - "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 16 cycles. During th= is period the front-end delivered no uops. Available PDIST counters: 0", + "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 16 cycles. During th= is period the front-end delivered no uops. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -287,7 +287,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_2", "MSRIndex": "0x3F7", "MSRValue": "0x600206", - "PublicDescription": "Retired instructions that are fetched after = an interval where the front-end delivered no uops for a period of at least = 2 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0", + "PublicDescription": "Retired instructions that are fetched after = an interval where the front-end delivered no uops for a period of at least = 2 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -299,7 +299,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_256", "MSRIndex": "0x3F7", "MSRValue": "0x610006", - "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 25= 6 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0", + "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 25= 6 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -311,7 +311,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_2_BUBBLES_GE_1", "MSRIndex": "0x3F7", "MSRValue": "0x100206", - "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after the front-end had at least 1 bubble-slot for a per= iod of 2 cycles. A bubble-slot is an empty issue-pipeline slot while there = was no RAT stall. Available PDIST counters: 0", + "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after the front-end had at least 1 bubble-slot for a per= iod of 2 cycles. A bubble-slot is an empty issue-pipeline slot while there = was no RAT stall. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -323,7 +323,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_32", "MSRIndex": "0x3F7", "MSRValue": "0x602006", - "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 32 cycles. During th= is period the front-end delivered no uops. Available PDIST counters: 0", + "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 32 cycles. During th= is period the front-end delivered no uops. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -335,7 +335,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_4", "MSRIndex": "0x3F7", "MSRValue": "0x600406", - "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 4 = cycles which was not interrupted by a back-end stall. Available PDIST count= ers: 0", + "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 4 = cycles which was not interrupted by a back-end stall. Available PDIST count= ers: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -347,7 +347,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_512", "MSRIndex": "0x3F7", "MSRValue": "0x620006", - "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 51= 2 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0", + "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 51= 2 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -359,7 +359,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_64", "MSRIndex": "0x3F7", "MSRValue": "0x604006", - "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 64= cycles which was not interrupted by a back-end stall. Available PDIST coun= ters: 0", + "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 64= cycles which was not interrupted by a back-end stall. Available PDIST coun= ters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -371,7 +371,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_8", "MSRIndex": "0x3F7", "MSRValue": "0x600806", - "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 8 cycles. During thi= s period the front-end delivered no uops. Available PDIST counters: 0", + "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 8 cycles. During thi= s period the front-end delivered no uops. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -383,7 +383,7 @@ "EventName": "FRONTEND_RETIRED.MISP_ANT", "MSRIndex": "0x3F7", "MSRValue": "0x9", - "PublicDescription": "ANT retired branches that got just mispredic= ted Available PDIST counters: 0", + "PublicDescription": "ANT retired branches that got just mispredic= ted Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x2", "Unit": "cpu_core" @@ -395,7 +395,7 @@ "EventName": "FRONTEND_RETIRED.MS_FLOWS", "MSRIndex": "0x3F7", "MSRValue": "0x8", - "PublicDescription": "Counts flows delivered by the Microcode Sequ= encer Available PDIST counters: 0", + "PublicDescription": "Counts flows delivered by the Microcode Sequ= encer Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -443,7 +443,7 @@ "EventName": "FRONTEND_RETIRED.STLB_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x15", - "PublicDescription": "Counts retired Instructions that experienced= STLB (2nd level TLB) true miss. Available PDIST counters: 0", + "PublicDescription": "Counts retired Instructions that experienced= STLB (2nd level TLB) true miss. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -455,7 +455,7 @@ "EventName": "FRONTEND_RETIRED.UNKNOWN_BRANCH", "MSRIndex": "0x3F7", "MSRValue": "0x17", - "PublicDescription": "Number retired branch instructions that caus= ed the front-end to be resteered when it finds the instruction in a fetch l= ine. This is called Unknown Branch which occurs for the first time a branch= instruction is fetched or when the branch is not tracked by the BPU (Branc= h Prediction Unit) anymore. Available PDIST counters: 0", + "PublicDescription": "Number retired branch instructions that caus= ed the front-end to be resteered when it finds the instruction in a fetch l= ine. This is called Unknown Branch which occurs for the first time a branch= instruction is fetched or when the branch is not tracked by the BPU (Branc= h Prediction Unit) anymore. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/arrowlake/memory.json b/tools/p= erf/pmu-events/arch/x86/arrowlake/memory.json index fb8d4ac69bda..1e6360347c0f 100644 --- a/tools/perf/pmu-events/arch/x86/arrowlake/memory.json +++ b/tools/perf/pmu-events/arch/x86/arrowlake/memory.json @@ -163,7 +163,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_1024", "MSRIndex": "0x3F6", "MSRValue": "0x400", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 1024 cycles. Reporte= d latency may be longer than just the memory latency. Available PDIST count= ers: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 1024 cycles. Reporte= d latency may be longer than just the memory latency.", "SampleAfterValue": "53", "UMask": "0x1", "Unit": "cpu_core" @@ -176,7 +176,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_128", "MSRIndex": "0x3F6", "MSRValue": "0x80", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 128 cycles. Reported= latency may be longer than just the memory latency. Available PDIST counte= rs: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 128 cycles. Reported= latency may be longer than just the memory latency.", "SampleAfterValue": "1009", "UMask": "0x1", "Unit": "cpu_core" @@ -189,7 +189,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_16", "MSRIndex": "0x3F6", "MSRValue": "0x10", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 16 cycles. Reported = latency may be longer than just the memory latency. Available PDIST counter= s: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 16 cycles. Reported = latency may be longer than just the memory latency.", "SampleAfterValue": "20011", "UMask": "0x1", "Unit": "cpu_core" @@ -202,7 +202,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_2048", "MSRIndex": "0x3F6", "MSRValue": "0x800", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 2048 cycles. Reporte= d latency may be longer than just the memory latency. Available PDIST count= ers: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 2048 cycles. Reporte= d latency may be longer than just the memory latency.", "SampleAfterValue": "23", "UMask": "0x1", "Unit": "cpu_core" @@ -215,7 +215,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_256", "MSRIndex": "0x3F6", "MSRValue": "0x100", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 256 cycles. Reported= latency may be longer than just the memory latency. Available PDIST counte= rs: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 256 cycles. Reported= latency may be longer than just the memory latency.", "SampleAfterValue": "503", "UMask": "0x1", "Unit": "cpu_core" @@ -228,7 +228,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_32", "MSRIndex": "0x3F6", "MSRValue": "0x20", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 32 cycles. Reported = latency may be longer than just the memory latency. Available PDIST counter= s: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 32 cycles. Reported = latency may be longer than just the memory latency.", "SampleAfterValue": "100007", "UMask": "0x1", "Unit": "cpu_core" @@ -241,7 +241,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_4", "MSRIndex": "0x3F6", "MSRValue": "0x4", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 4 cycles. Reported l= atency may be longer than just the memory latency. Available PDIST counters= : 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 4 cycles. Reported l= atency may be longer than just the memory latency.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -254,7 +254,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_512", "MSRIndex": "0x3F6", "MSRValue": "0x200", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 512 cycles. Reported= latency may be longer than just the memory latency. Available PDIST counte= rs: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 512 cycles. Reported= latency may be longer than just the memory latency.", "SampleAfterValue": "101", "UMask": "0x1", "Unit": "cpu_core" @@ -267,7 +267,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_64", "MSRIndex": "0x3F6", "MSRValue": "0x40", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 64 cycles. Reported = latency may be longer than just the memory latency. Available PDIST counter= s: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 64 cycles. Reported = latency may be longer than just the memory latency.", "SampleAfterValue": "2003", "UMask": "0x1", "Unit": "cpu_core" @@ -280,7 +280,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_8", "MSRIndex": "0x3F6", "MSRValue": "0x8", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 8 cycles. Reported l= atency may be longer than just the memory latency. Available PDIST counters= : 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 8 cycles. Reported l= atency may be longer than just the memory latency.", "SampleAfterValue": "50021", "UMask": "0x1", "Unit": "cpu_core" @@ -291,7 +291,7 @@ "Data_LA": "1", "EventCode": "0xcd", "EventName": "MEM_TRANS_RETIRED.STORE_SAMPLE", - "PublicDescription": "Counts Retired memory accesses with at least= 1 store operation. This PEBS event is the precisely-distributed (PDist) tr= igger covering all stores uops for sampling by the PEBS Store Latency Facil= ity. The facility is described in Intel SDM Volume 3 section 19.9.8 Availab= le PDIST counters: 0", + "PublicDescription": "Counts Retired memory accesses with at least= 1 store operation. This PEBS event is the precisely-distributed (PDist) tr= igger covering all stores uops for sampling by the PEBS Store Latency Facil= ity. The facility is described in Intel SDM Volume 3 section 19.9.8 Availab= le PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/arrowlake/pipeline.json b/tools= /perf/pmu-events/arch/x86/arrowlake/pipeline.json index 18a22368b99b..0651e2c4561e 100644 --- a/tools/perf/pmu-events/arch/x86/arrowlake/pipeline.json +++ b/tools/perf/pmu-events/arch/x86/arrowlake/pipeline.json @@ -74,13 +74,14 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.ALL_BRANCHES", - "PublicDescription": "Counts all branch instructions retired. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts all branch instructions retired. Avai= lable PDIST counters: 0,1", "SampleAfterValue": "400009", "Unit": "cpu_core" }, { "BriefDescription": "Counts the total number of branch instruction= s retired for all branch types.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "ARL010, ARL011", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.ALL_BRANCHES", "PublicDescription": "Counts the total number of instructions in w= hich the instruction pointer (IP) of the processor is resteered due to a br= anch instruction and the branch instruction successfully retires. All bran= ch type instructions are accounted for.", @@ -101,7 +102,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND", - "PublicDescription": "Counts conditional branch instructions retir= ed. Available PDIST counters: 0", + "PublicDescription": "Counts conditional branch instructions retir= ed. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x111", "Unit": "cpu_core" @@ -109,6 +110,7 @@ { "BriefDescription": "Counts the number of retired JCC (Jump on Con= ditional Code) branch instructions retired, includes both taken and not tak= en branches.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "ARL011", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND", "SampleAfterValue": "200003", @@ -120,7 +122,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND_NTAKEN", - "PublicDescription": "Counts not taken branch instructions retired= . Available PDIST counters: 0", + "PublicDescription": "Counts not taken branch instructions retired= . Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x10", "Unit": "cpu_core" @@ -139,7 +141,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND_TAKEN", - "PublicDescription": "Counts taken conditional branch instructions= retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken conditional branch instructions= retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x101", "Unit": "cpu_core" @@ -158,7 +160,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND_TAKEN_BWD", - "PublicDescription": "Counts taken backward conditional branch ins= tructions retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken backward conditional branch ins= tructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x1", "Unit": "cpu_core" @@ -168,7 +170,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND_TAKEN_FWD", - "PublicDescription": "Counts taken forward conditional branch inst= ructions retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken forward conditional branch inst= ructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x102", "Unit": "cpu_core" @@ -187,7 +189,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.FAR_BRANCH", - "PublicDescription": "Counts far branch instructions retired. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts far branch instructions retired. Avai= lable PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x40", "Unit": "cpu_core" @@ -195,6 +197,7 @@ { "BriefDescription": "Counts the number of far branch instructions = retired, includes far jump, far call and return, and interrupt call and ret= urn.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "ARL011", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.FAR_BRANCH", "SampleAfterValue": "200003", @@ -215,7 +218,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.INDIRECT", - "PublicDescription": "Counts near indirect branch instructions ret= ired excluding returns. TSX abort is an indirect branch. Available PDIST co= unters: 0", + "PublicDescription": "Counts near indirect branch instructions ret= ired excluding returns. TSX abort is an indirect branch. Available PDIST co= unters: 0,1", "SampleAfterValue": "100003", "UMask": "0x80", "Unit": "cpu_core" @@ -223,6 +226,7 @@ { "BriefDescription": "Counts the number of near indirect JMP and ne= ar indirect CALL branch instructions retired.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "ARL011", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.INDIRECT", "SampleAfterValue": "200003", @@ -241,6 +245,7 @@ { "BriefDescription": "Counts the number of near indirect CALL branc= h instructions retired.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "ARL011", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.INDIRECT_CALL", "SampleAfterValue": "200003", @@ -270,7 +275,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_CALL", - "PublicDescription": "Counts both direct and indirect near call in= structions retired. Available PDIST counters: 0", + "PublicDescription": "Counts both direct and indirect near call in= structions retired. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x2", "Unit": "cpu_core" @@ -278,6 +283,7 @@ { "BriefDescription": "Counts the number of near CALL branch instruc= tions retired.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "ARL010, ARL011", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_CALL", "SampleAfterValue": "200003", @@ -298,7 +304,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_RETURN", - "PublicDescription": "Counts return instructions retired. Availabl= e PDIST counters: 0", + "PublicDescription": "Counts return instructions retired. Availabl= e PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x8", "Unit": "cpu_core" @@ -317,7 +323,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_TAKEN", - "PublicDescription": "Counts taken branch instructions retired. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts taken branch instructions retired. Av= ailable PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x20", "Unit": "cpu_core" @@ -325,6 +331,7 @@ { "BriefDescription": "Counts the number of near taken branch instru= ctions retired.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "ARL011", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_TAKEN", "SampleAfterValue": "200003", @@ -372,7 +379,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.ALL_BRANCHES", - "PublicDescription": "Counts all the retired branch instructions t= hat were mispredicted by the processor. A branch misprediction occurs when = the processor incorrectly predicts the destination of the branch. When the= misprediction is discovered at execution, all the instructions executed in= the wrong (speculative) path must be discarded, and the processor must sta= rt fetching from the correct path. Available PDIST counters: 0", + "PublicDescription": "Counts all the retired branch instructions t= hat were mispredicted by the processor. A branch misprediction occurs when = the processor incorrectly predicts the destination of the branch. When the= misprediction is discovered at execution, all the instructions executed in= the wrong (speculative) path must be discarded, and the processor must sta= rt fetching from the correct path. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "Unit": "cpu_core" }, @@ -390,7 +397,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.ALL_BRANCHES_COST", - "PublicDescription": "All mispredicted branch instructions retired= . This precise event may be used to get the misprediction cost via the Reti= re_Latency field of PEBS. It fires on the instruction that immediately foll= ows the mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "All mispredicted branch instructions retired= . This precise event may be used to get the misprediction cost via the Reti= re_Latency field of PEBS. It fires on the instruction that immediately foll= ows the mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x44", "Unit": "cpu_core" @@ -409,7 +416,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND", - "PublicDescription": "Counts mispredicted conditional branch instr= uctions retired. Available PDIST counters: 0", + "PublicDescription": "Counts mispredicted conditional branch instr= uctions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x111", "Unit": "cpu_core" @@ -428,7 +435,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_COST", - "PublicDescription": "Mispredicted conditional branch instructions= retired. This precise event may be used to get the misprediction cost via = the Retire_Latency field of PEBS. It fires on the instruction that immediat= ely follows the mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted conditional branch instructions= retired. This precise event may be used to get the misprediction cost via = the Retire_Latency field of PEBS. It fires on the instruction that immediat= ely follows the mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x151", "Unit": "cpu_core" @@ -438,7 +445,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_NTAKEN", - "PublicDescription": "Counts the number of conditional branch inst= ructions retired that were mispredicted and the branch direction was not ta= ken. Available PDIST counters: 0", + "PublicDescription": "Counts the number of conditional branch inst= ructions retired that were mispredicted and the branch direction was not ta= ken. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x10", "Unit": "cpu_core" @@ -448,7 +455,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_NTAKEN_COST", - "PublicDescription": "Mispredicted non-taken conditional branch in= structions retired. This precise event may be used to get the misprediction= cost via the Retire_Latency field of PEBS. It fires on the instruction tha= t immediately follows the mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted non-taken conditional branch in= structions retired. This precise event may be used to get the misprediction= cost via the Retire_Latency field of PEBS. It fires on the instruction tha= t immediately follows the mispredicted branch. Available PDIST counters: 0,= 1", "SampleAfterValue": "400009", "UMask": "0x50", "Unit": "cpu_core" @@ -467,7 +474,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN", - "PublicDescription": "Counts taken conditional mispredicted branch= instructions retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken conditional mispredicted branch= instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x101", "Unit": "cpu_core" @@ -486,7 +493,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN_BWD", - "PublicDescription": "Counts taken backward conditional mispredict= ed branch instructions retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken backward conditional mispredict= ed branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x1", "Unit": "cpu_core" @@ -496,7 +503,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN_BWD_COST", - "PublicDescription": "number of branch instructions retired that w= ere mispredicted and taken backward. This precise event may be used to get = the misprediction cost via the Retire_Latency field of PEBS. It fires on th= e instruction that immediately follows the mispredicted branch. Available P= DIST counters: 0", + "PublicDescription": "number of branch instructions retired that w= ere mispredicted and taken backward. This precise event may be used to get = the misprediction cost via the Retire_Latency field of PEBS. It fires on th= e instruction that immediately follows the mispredicted branch. Available P= DIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x8001", "Unit": "cpu_core" @@ -506,7 +513,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN_COST", - "PublicDescription": "Mispredicted taken conditional branch instru= ctions retired. This precise event may be used to get the misprediction cos= t via the Retire_Latency field of PEBS. It fires on the instruction that im= mediately follows the mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted taken conditional branch instru= ctions retired. This precise event may be used to get the misprediction cos= t via the Retire_Latency field of PEBS. It fires on the instruction that im= mediately follows the mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x141", "Unit": "cpu_core" @@ -516,7 +523,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN_FWD", - "PublicDescription": "Counts taken forward conditional mispredicte= d branch instructions retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken forward conditional mispredicte= d branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "Unit": "cpu_core" }, @@ -525,7 +532,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN_FWD_COST", - "PublicDescription": "number of branch instructions retired that w= ere mispredicted and taken forward. This precise event may be used to get t= he misprediction cost via the Retire_Latency field of PEBS. It fires on the= instruction that immediately follows the mispredicted branch. Available PD= IST counters: 0", + "PublicDescription": "number of branch instructions retired that w= ere mispredicted and taken forward. This precise event may be used to get t= he misprediction cost via the Retire_Latency field of PEBS. It fires on the= instruction that immediately follows the mispredicted branch. Available PD= IST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x8002", "Unit": "cpu_core" @@ -544,7 +551,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.INDIRECT", - "PublicDescription": "Counts miss-predicted near indirect branch i= nstructions retired excluding returns. TSX abort is an indirect branch. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts miss-predicted near indirect branch i= nstructions retired excluding returns. TSX abort is an indirect branch. Ava= ilable PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x80", "Unit": "cpu_core" @@ -572,7 +579,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.INDIRECT_CALL", - "PublicDescription": "Counts retired mispredicted indirect (near t= aken) CALL instructions, including both register and memory indirect. Avail= able PDIST counters: 0", + "PublicDescription": "Counts retired mispredicted indirect (near t= aken) CALL instructions, including both register and memory indirect. Avail= able PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x2", "Unit": "cpu_core" @@ -591,7 +598,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.INDIRECT_CALL_COST", - "PublicDescription": "Mispredicted indirect CALL retired. This pre= cise event may be used to get the misprediction cost via the Retire_Latency= field of PEBS. It fires on the instruction that immediately follows the mi= spredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted indirect CALL retired. This pre= cise event may be used to get the misprediction cost via the Retire_Latency= field of PEBS. It fires on the instruction that immediately follows the mi= spredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x42", "Unit": "cpu_core" @@ -601,7 +608,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.INDIRECT_COST", - "PublicDescription": "Mispredicted near indirect branch instructio= ns retired (excluding returns). This precise event may be used to get the m= isprediction cost via the Retire_Latency field of PEBS. It fires on the ins= truction that immediately follows the mispredicted branch. Available PDIST = counters: 0", + "PublicDescription": "Mispredicted near indirect branch instructio= ns retired (excluding returns). This precise event may be used to get the m= isprediction cost via the Retire_Latency field of PEBS. It fires on the ins= truction that immediately follows the mispredicted branch. Available PDIST = counters: 0,1", "SampleAfterValue": "100003", "UMask": "0xc0", "Unit": "cpu_core" @@ -620,7 +627,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.NEAR_TAKEN", - "PublicDescription": "Counts number of near branch instructions re= tired that were mispredicted and taken. Available PDIST counters: 0", + "PublicDescription": "Counts number of near branch instructions re= tired that were mispredicted and taken. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x20", "Unit": "cpu_core" @@ -639,7 +646,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.NEAR_TAKEN_COST", - "PublicDescription": "Mispredicted taken near branch instructions = retired. This precise event may be used to get the misprediction cost via t= he Retire_Latency field of PEBS. It fires on the instruction that immediate= ly follows the mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted taken near branch instructions = retired. This precise event may be used to get the misprediction cost via t= he Retire_Latency field of PEBS. It fires on the instruction that immediate= ly follows the mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x60", "Unit": "cpu_core" @@ -649,7 +656,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.RET", - "PublicDescription": "This is a non-precise version (that is, does= not use PEBS) of the event that counts mispredicted return instructions re= tired. Available PDIST counters: 0", + "PublicDescription": "This is a non-precise version (that is, does= not use PEBS) of the event that counts mispredicted return instructions re= tired. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x8", "Unit": "cpu_core" @@ -677,7 +684,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.RET_COST", - "PublicDescription": "Mispredicted ret instructions retired. This = precise event may be used to get the misprediction cost via the Retire_Late= ncy field of PEBS. It fires on the instruction that immediately follows the= mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted ret instructions retired. This = precise event may be used to get the misprediction cost via the Retire_Late= ncy field of PEBS. It fires on the instruction that immediately follows the= mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x48", "Unit": "cpu_core" @@ -1046,7 +1053,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc0", "EventName": "INST_RETIRED.ANY_P", - "PublicDescription": "Counts the number of X86 instructions retire= d - an Architectural PerfMon event. Counting continues during hardware inte= rrupts, traps, and inside interrupt handlers. Notes: INST_RETIRED.ANY is co= unted by a designated fixed counter freeing up programmable counters to cou= nt other events. INST_RETIRED.ANY_P is counted by a programmable counter. A= vailable PDIST counters: 0", + "PublicDescription": "Counts the number of X86 instructions retire= d - an Architectural PerfMon event. Counting continues during hardware inte= rrupts, traps, and inside interrupt handlers. Notes: INST_RETIRED.ANY is co= unted by a designated fixed counter freeing up programmable counters to cou= nt other events. INST_RETIRED.ANY_P is counted by a programmable counter. A= vailable PDIST counters: 0,1", "SampleAfterValue": "2000003", "Unit": "cpu_core" }, @@ -1063,7 +1070,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc0", "EventName": "INST_RETIRED.BR_FUSED", - "PublicDescription": "retired macro-fused uops when there is a bra= nch in the macro-fused pair (the two instructions that got macro-fused coun= t once in this pmon) Available PDIST counters: 0", + "PublicDescription": "retired macro-fused uops when there is a bra= nch in the macro-fused pair (the two instructions that got macro-fused coun= t once in this pmon) Available PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -1073,7 +1080,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc0", "EventName": "INST_RETIRED.MACRO_FUSED", - "PublicDescription": "INST_RETIRED.MACRO_FUSED Available PDIST cou= nters: 0", + "PublicDescription": "INST_RETIRED.MACRO_FUSED Available PDIST cou= nters: 0,1", "SampleAfterValue": "2000003", "UMask": "0x30", "Unit": "cpu_core" @@ -1083,7 +1090,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc0", "EventName": "INST_RETIRED.NOP", - "PublicDescription": "Counts all retired NOP or ENDBR32/64 or PREF= ETCHIT0/1 instructions Available PDIST counters: 0", + "PublicDescription": "Counts all retired NOP or ENDBR32/64 or PREF= ETCHIT0/1 instructions Available PDIST counters: 0,1", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1102,7 +1109,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc0", "EventName": "INST_RETIRED.REP_ITERATION", - "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent. Available PDIST counters: 0", + "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent. Available PDIST counters: 0,1", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -1348,6 +1355,15 @@ "UMask": "0x88", "Unit": "cpu_core" }, + { + "BriefDescription": "Counts the number of times a load got early b= locked due to preceding store operation with unknown address or unknown dat= a. Excluding in-line (immediate) wakeups", + "Counter": "0,1,2,3,4,5,6,7,8,9", + "EventCode": "0x03", + "EventName": "LD_BLOCKS.STORE_EARLY", + "SampleAfterValue": "100003", + "UMask": "0xa1", + "Unit": "cpu_core" + }, { "BriefDescription": "Counts the number of occurrences a retired lo= ad gets blocked because its address partially overlaps with an older store = (size mismatch) - unknown_sta/bad_forward", "Counter": "0,1,2,3,4,5,6,7", @@ -1563,7 +1579,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xe4", "EventName": "MISC_RETIRED.LBR_INSERTS", - "PublicDescription": "LBR record is inserted Available PDIST count= ers: 0", + "PublicDescription": "LBR record is inserted Available PDIST count= ers: 0,1", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1929,7 +1945,7 @@ }, { "BriefDescription": "Fixed Counter: Counts the number of retiremen= t slots not consumed due to front end stalls.", - "Counter": "37", + "Counter": "Fixed counter 5", "EventName": "TOPDOWN_FE_BOUND.ALL", "SampleAfterValue": "1000003", "UMask": "0x6", @@ -2126,7 +2142,7 @@ }, { "BriefDescription": "Fixed Counter: Counts the number of consumed = retirement slots.", - "Counter": "38", + "Counter": "Fixed counter 6", "EventName": "TOPDOWN_RETIRING.ALL", "SampleAfterValue": "1000003", "UMask": "0x7", diff --git a/tools/perf/pmu-events/arch/x86/mapfile.csv b/tools/perf/pmu-ev= ents/arch/x86/mapfile.csv index d0a17905c74e..12017f568606 100644 --- a/tools/perf/pmu-events/arch/x86/mapfile.csv +++ b/tools/perf/pmu-events/arch/x86/mapfile.csv @@ -1,7 +1,7 @@ Family-model,Version,Filename,EventType GenuineIntel-6-(97|9A|B7|BA|BF),v1.33,alderlake,core GenuineIntel-6-BE,v1.33,alderlaken,core -GenuineIntel-6-C[56],v1.09,arrowlake,core +GenuineIntel-6-C[56],v1.12,arrowlake,core GenuineIntel-6-(1C|26|27|35|36),v5,bonnell,core GenuineIntel-6-(3D|47),v30,broadwell,core GenuineIntel-6-56,v12,broadwellde,core --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pg1-f201.google.com (mail-pg1-f201.google.com [209.85.215.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 6A3BF273808 for ; Mon, 18 Aug 2025 19:05:29 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.215.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543931; cv=none; b=OSvXtzqdqJiKKJrXUbiBx+fI7rhc5IA837l5ILJwMuSMdzxna0KVngGgWAv7mvMJn0WLYIxBxJmCGTK4wKsVjoNVaY237jN9V9Cfu+99rM5jrIK4G+nlzr6fBNt6wnXz6dTLwGk7xYZdzTCVs9z4uOpiu87p8ynlLLOZelmYA3I= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543931; c=relaxed/simple; bh=9S9R1Tnr+HLClnwrwFBXoP9U/necgxJT61Te3T67sOo=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=emxucOsuraKTjLog1PFQ5HcZKzBjvj825+Ouz5LCJiNT+8BHyWDiB20owZNDvTfAMn5eYNg1dcGdk6s/mPhP5eQLmGt9+btvd0r++CyPlWbcVtsaLBU+0Zwg1Mt41eKRTzla52QXsNyibPlV4Zzm/d3DG1KB/SdKwq8bUQOwIBY= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=flNJ5Pj2; arc=none smtp.client-ip=209.85.215.201 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="flNJ5Pj2" Received: by mail-pg1-f201.google.com with SMTP id 41be03b00d2f7-b474b68cff7so924158a12.1 for ; Mon, 18 Aug 2025 12:05:29 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543929; x=1756148729; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=O8jNYADsYAC3I3NWtY+vdwe29CzDwhW0Y4Q7x0BOi5w=; b=flNJ5Pj2YRodhVCjwMz3W5/YvOA/WHFTFPPYzJxo8Pi8UIryvAT9XtTe+9WYtZyUJS mOVLF1+b3qgOTdQ3WvxbsTREugyHlnp8HErM4iILA9fv3Dtsf/HzqtdUrjLjcscWVNOB 8dsFtKYOpCXxhCwFJkxuAe2dQtSi/xgCwGGLRy3MomMrp9g03oG+9GneHshUk9++hucx yTfjtHcqTAjU0XD7VnC2QBm2pejY79KQ8BN+eQOG9PJAaFOuf2xbqR3ZBHnTcX2NZEkN Jt0aPQpNrVPMytY8JHM++5UT6D9AcYCbYkW4COGUsIx7y49+wQrYdZcU9anC8j/xXjI6 kvuA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543929; x=1756148729; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=O8jNYADsYAC3I3NWtY+vdwe29CzDwhW0Y4Q7x0BOi5w=; b=irTZj5F8/18Wy0H/GE2X3a/5P4a3jXYK3s5NMjlAQdJybjgpgRI5c5+5F6se39dfXU Mig9Aait0H3zO7mhpfwP/JNVczjATajpIV44wQzSMHQI9jM8EhVqlkz5vuGt6YRu+VVt ZdTufaIA/VGC9X1m91W4HMQAhtgRDu4Q7UjO8GUaY74ImwJhybAyxldcg3utmZDRocwX i6yhssBc+hydXxqQl3KdHqDWN1vo84iQOqNg2LamnQ9WHaYEqcuANi56GPbF6OICcpMN tR0BNpq2q5/flBbmA7Ge6jgbsUBJMYNYaCDyemY84w0oAVDsOd/enQMM/pt1qIxMM8mw STig== X-Forwarded-Encrypted: i=1; AJvYcCXayNbbqU/0noS3ecQSF9iXWlsTuSN7Lk+H/JzjP5SX0ObaBrqJTxQxo+qh0T+oj1iYfx0QIE/i2kIXztM=@vger.kernel.org X-Gm-Message-State: AOJu0YxQoouhQuLIH1dA29K25NXG9U/fgMBPmrwMURFD09hVbfRfIGPl piIGj/4UWQ1Ihxl8125o7LGfEDSv+/4wHFNYTQBWSd3srI21ZIs5XU+BpbYn8zUvz+6FcRFvtpD njmT/wBGiCw== X-Google-Smtp-Source: AGHT+IHloI8Nn8mfI9y+rZjcc+JEPkbdPm7VxDTfRcR3DJNj0pzKGCy2/i9KOSh4IXr6QqWDJyfmr1nK/dKD X-Received: from pjbpq4.prod.google.com ([2002:a17:90b:3d84:b0:30a:7da4:f075]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:90b:524d:b0:31f:336a:f0e1 with SMTP id 98e67ed59e1d1-3237f5b41c9mr507831a91.17.1755543928665; Mon, 18 Aug 2025 12:05:28 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:00 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-5-irogers@google.com> Subject: [PATCH v2 04/20] perf vendor events: Update broadwell metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/broadwell/bdw-metrics.json | 30 ++++++++--------- .../arch/x86/broadwellde/bdwde-metrics.json | 30 ++++++++--------- .../arch/x86/broadwellx/bdx-metrics.json | 33 +++++++++---------- 3 files changed, 42 insertions(+), 51 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/broadwell/bdw-metrics.json b/to= ols/perf/pmu-events/arch/x86/broadwell/bdw-metrics.json index 89750117a7f6..1d8e910f5961 100644 --- a/tools/perf/pmu-events/arch/x86/broadwell/bdw-metrics.json +++ b/tools/perf/pmu-events/arch/x86/broadwell/bdw-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -80,7 +80,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_0 + UOPS_DISPATCHED_PORT= .PORT_1 + UOPS_DISPATCHED_PORT.PORT_5 + UOPS_DISPATCHED_PORT.PORT_6) / tma_= info_thread_slots", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_alu_op_utilization", @@ -98,7 +97,6 @@ }, { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma= _retiring)", "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", @@ -139,7 +137,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU retired uops originated from CISC (complex instruction set computer) in= struction", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "max(0, tma_microcode_sequencer - tma_assists)", "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_gro= up", "MetricName": "tma_cisc", @@ -640,7 +637,7 @@ "MetricThreshold": "tma_info_memory_tlb_page_walks_utilization > 0= .5" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (cpu@UOPS_EXECUTED.CORE\\,cm= ask\\=3D1@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -653,7 +650,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -665,7 +662,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -854,7 +851,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution port for Load operations", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_2 + UOPS_DISPATCHED_PORT= .PORT_3 + UOPS_DISPATCHED_PORT.PORT_7 - UOPS_DISPATCHED_PORT.PORT_4) / (2 *= tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_load_op_utilization", @@ -1032,7 +1028,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", - "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else (CYCLE_ACTIVITY.STALLS_TOTAL - (RS_EVENTS.EMPTY_CYCLES if tma= _fetch_latency > 0.1 else 0)) / tma_info_core_core_clks)", + "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else CYCLE_ACTIVITY.STALLS_TOTAL - (RS_EVENTS.EMPTY_CYCLES if tma_= fetch_latency > 0.1 else 0)) / tma_info_core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", "MetricThreshold": "tma_ports_utilized_0 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1041,7 +1037,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else (UOPS_EXECUTED.CYCLES_GE_= 1_UOP_EXEC - UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC) / tma_info_core_core_clks= )", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1= _UOP_EXEC - UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC) / tma_info_core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", "MetricThreshold": "tma_ports_utilized_1 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1050,7 +1046,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else (UOPS_EXECUTED.CYCLES_GE_= 2_UOPS_EXEC - UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clk= s)", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_2= _UOPS_EXEC - UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clks= ", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", "MetricThreshold": "tma_ports_utilized_2 > 0.15 & (tma_ports_utili= zation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", diff --git a/tools/perf/pmu-events/arch/x86/broadwellde/bdwde-metrics.json = b/tools/perf/pmu-events/arch/x86/broadwellde/bdwde-metrics.json index 81175f0f2603..a5e408ca46a7 100644 --- a/tools/perf/pmu-events/arch/x86/broadwellde/bdwde-metrics.json +++ b/tools/perf/pmu-events/arch/x86/broadwellde/bdwde-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -80,7 +80,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_0 + UOPS_DISPATCHED_PORT= .PORT_1 + UOPS_DISPATCHED_PORT.PORT_5 + UOPS_DISPATCHED_PORT.PORT_6) / tma_= info_thread_slots", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_alu_op_utilization", @@ -98,7 +97,6 @@ }, { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma= _retiring)", "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", @@ -139,7 +137,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU retired uops originated from CISC (complex instruction set computer) in= struction", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "max(0, tma_microcode_sequencer - tma_assists)", "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_gro= up", "MetricName": "tma_cisc", @@ -632,7 +629,7 @@ "MetricThreshold": "tma_info_memory_tlb_page_walks_utilization > 0= .5" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (cpu@UOPS_EXECUTED.CORE\\,cm= ask\\=3D1@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -645,7 +642,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -657,7 +654,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -846,7 +843,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution port for Load operations", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_2 + UOPS_DISPATCHED_PORT= .PORT_3 + UOPS_DISPATCHED_PORT.PORT_7 - UOPS_DISPATCHED_PORT.PORT_4) / (2 *= tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_load_op_utilization", @@ -1021,7 +1017,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", - "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else (CYCLE_ACTIVITY.STALLS_TOTAL - (RS_EVENTS.EMPTY_CYCLES if tma= _fetch_latency > 0.1 else 0)) / tma_info_core_core_clks)", + "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else CYCLE_ACTIVITY.STALLS_TOTAL - (RS_EVENTS.EMPTY_CYCLES if tma_= fetch_latency > 0.1 else 0)) / tma_info_core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", "MetricThreshold": "tma_ports_utilized_0 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1030,7 +1026,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else (UOPS_EXECUTED.CYCLES_GE_= 1_UOP_EXEC - UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC) / tma_info_core_core_clks= )", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1= _UOP_EXEC - UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC) / tma_info_core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", "MetricThreshold": "tma_ports_utilized_1 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1039,7 +1035,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else (UOPS_EXECUTED.CYCLES_GE_= 2_UOPS_EXEC - UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clk= s)", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_2= _UOPS_EXEC - UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clks= ", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", "MetricThreshold": "tma_ports_utilized_2 > 0.15 & (tma_ports_utili= zation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", diff --git a/tools/perf/pmu-events/arch/x86/broadwellx/bdx-metrics.json b/t= ools/perf/pmu-events/arch/x86/broadwellx/bdx-metrics.json index 5d06a3f72be2..5b83b040060c 100644 --- a/tools/perf/pmu-events/arch/x86/broadwellx/bdx-metrics.json +++ b/tools/perf/pmu-events/arch/x86/broadwellx/bdx-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -282,7 +282,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_0 + UOPS_DISPATCHED_PORT= .PORT_1 + UOPS_DISPATCHED_PORT.PORT_5 + UOPS_DISPATCHED_PORT.PORT_6) / tma_= info_thread_slots", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_alu_op_utilization", @@ -300,7 +299,6 @@ }, { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma= _retiring)", "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", @@ -341,7 +339,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU retired uops originated from CISC (complex instruction set computer) in= struction", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "max(0, tma_microcode_sequencer - tma_assists)", "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_gro= up", "MetricName": "tma_cisc", @@ -842,7 +839,7 @@ "MetricThreshold": "tma_info_memory_tlb_page_walks_utilization > 0= .5" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (cpu@UOPS_EXECUTED.CORE\\,cm= ask\\=3D1@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -855,7 +852,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -867,7 +864,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -907,6 +904,7 @@ }, { "BriefDescription": "Average number of parallel data read requests= to external memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "UNC_C_TOR_OCCUPANCY.MISS_OPCODE@filter_opc\\=3D0x18= 2@ / UNC_C_TOR_OCCUPANCY.MISS_OPCODE@filter_opc\\=3D0x182\\,thresh\\=3D1@", "MetricGroup": "Mem;MemoryBW;SoC", "MetricName": "tma_info_system_mem_parallel_reads", @@ -1076,7 +1074,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution port for Load operations", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_2 + UOPS_DISPATCHED_PORT= .PORT_3 + UOPS_DISPATCHED_PORT.PORT_7 - UOPS_DISPATCHED_PORT.PORT_4) / (2 *= tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_load_op_utilization", @@ -1086,6 +1083,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling loads from local memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "200 * (MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM * (= 1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOA= D_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UO= PS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_= LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE= _DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM + MEM_LOAD_UOPS_L3_MISS_R= ETIRED.REMOTE_FWD))) / tma_info_thread_clks", "MetricGroup": "Server;TopdownL5;tma_L5_group;tma_mem_latency_grou= p", "MetricName": "tma_local_mem", @@ -1263,7 +1261,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", - "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else (CYCLE_ACTIVITY.STALLS_TOTAL - (RS_EVENTS.EMPTY_CYCLES if tma= _fetch_latency > 0.1 else 0)) / tma_info_core_core_clks)", + "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else CYCLE_ACTIVITY.STALLS_TOTAL - (RS_EVENTS.EMPTY_CYCLES if tma_= fetch_latency > 0.1 else 0)) / tma_info_core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", "MetricThreshold": "tma_ports_utilized_0 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1272,7 +1270,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else (UOPS_EXECUTED.CYCLES_GE_= 1_UOP_EXEC - UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC) / tma_info_core_core_clks= )", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1= _UOP_EXEC - UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC) / tma_info_core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", "MetricThreshold": "tma_ports_utilized_1 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1281,7 +1279,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else (UOPS_EXECUTED.CYCLES_GE_= 2_UOPS_EXEC - UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clk= s)", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_2= _UOPS_EXEC - UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clks= ", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", "MetricThreshold": "tma_ports_utilized_2 > 0.15 & (tma_ports_utili= zation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1308,6 +1306,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling loads from remote memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "310 * (MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_DRAM * = (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LO= AD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_U= OPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM= _LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOT= E_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM + MEM_LOAD_UOPS_L3_MISS_= RETIRED.REMOTE_FWD))) / tma_info_thread_clks", "MetricGroup": "Server;Snoop;TopdownL5;tma_L5_group;tma_mem_latenc= y_group", "MetricName": "tma_remote_mem", --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pg1-f201.google.com (mail-pg1-f201.google.com [209.85.215.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 7575F3451A1 for ; Mon, 18 Aug 2025 19:05:31 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.215.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543934; cv=none; b=OshvrVawN+9mKG0Hx3uI+XEvueMesy2OboJk0iOkfOFEX8c2KOhwqKGEA1Fyx+gfHYyxL5xdyYZihgJXk6RW8h+imf20D71F1oFV81J5TbGBKiY/w+waIbQW01LfXPdjnIPdSqcWBirQXNasI04RS2G1G2AGxE/Ok8bMp8kG62A= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543934; c=relaxed/simple; bh=1Mxn96s2JkXeYuxz3dn41Yj2k6d9mCQ/rvwLvu99xTs=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=kZ1P3ntx/+gNnTJv+WP3ayEx3mXK+j45XDkIml6crm4b/ip9W7ViHp/K8KNIiLJS87DDR7ZbFZS94aAoVylqP0J6FWhVLCunOnBLf+OBG+CwtZ/d70Idy5pn0d96IJ0KtHJguxPP+I363NCwh7NVMm/c1095+L9v1qxDxBIheLs= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=KfNfe4qM; arc=none smtp.client-ip=209.85.215.201 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="KfNfe4qM" Received: by mail-pg1-f201.google.com with SMTP id 41be03b00d2f7-b4716fc56a9so7538823a12.0 for ; Mon, 18 Aug 2025 12:05:31 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543931; x=1756148731; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=DDtALuobt3+sYWvXnEUnbCSjfCAbVdakwJdFNMGEAg0=; b=KfNfe4qMAN6IayO2qG+GsViRjXlbNUn+GHyXnjcs4xH6k6oBAIaqKMncXo3wrfGeXo QtEER0u22QHDRvJ0ur6mnpaKkVkiRDzUbpICHYCCDOoizn3PnBwWRhLUdvSvXzAFOOQZ TfuXeGULxz5jEvmm6/UZcW+vfIs76jHU84m4a4xWffT5l0kPv+CN/PbxVlTLhdUTuxn1 3hOFZ1WNl11OsoZyzxpnL74swrYJX5ncVaBdSzOrsb/RnMXYmJAojA49eCfb9A/NknLM G7Pj23hr2eWYOfA/fU2ORDkEz7Xgbp/x+nuoYvx6InE0XdKICrkbeynTua9tif6+iNDr FQag== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543931; x=1756148731; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=DDtALuobt3+sYWvXnEUnbCSjfCAbVdakwJdFNMGEAg0=; b=PAe9S6fp6faDVZ/+0k7sYZ2cxF3sucgQ0wlvMcPSuddufjSAlfOfu4xK3OjRf23gSG F8LStq9q5I5lkQ+oc1y6aK6AaeP3E8tMjhoh1tZF3dOoKlHbrTLjYpAaDNrUbMvcDEJv 6QLfzvhBMjPRRGpQnUjis9yqZeCzp8qcH+RijFPqTwpBNG2/UxpSCxyr8PLrGZB+EYnk n9oI2hKczvz49dniLV8ArvpIlYTP9bssGxeL2z3e22b05EJTxqYj2d8cvMVVGKKFQPWK DQnBBZw5iPJdUuNsqfZOmxYorIyHvmCSY7tIzE8pP3RFHEKmmfIHEceDQM5MuCQdDXp+ cA9w== X-Forwarded-Encrypted: i=1; AJvYcCWG4suaQrYAbMIjcerRPQT1QGE8WQKWfWN/1BTCx8iuYuYTLSB/iGKue1liFfRz53RO0733EFerYZ6iqZ4=@vger.kernel.org X-Gm-Message-State: AOJu0YyldCXk0PGjPaWMSPjVy3iudaqGWpS65XaJwyFLTzrr41tCq96f XrjySrpEigH7Bcrw7Ch1o8pBOJKmElX9yFgul26+PwMC9U0jZlEmVXMYBX/Wh2ir7/jHdTsw6vJ hko/NV5g+ww== X-Google-Smtp-Source: AGHT+IEecladAs2jHw1rKrcGB6j+E5XGEQRkYs8RI9oKrEk8Em2jq00XMBdW5rgDjQfk2qMHv3VrMlj5PGTN X-Received: from plik6.prod.google.com ([2002:a17:903:3846:b0:23f:fded:852b]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:903:1a26:b0:240:52c8:2556 with SMTP id d9443c01a7336-2446d95ab7emr231073395ad.39.1755543930691; Mon, 18 Aug 2025 12:05:30 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:01 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-6-irogers@google.com> Subject: [PATCH v2 05/20] perf vendor events: Update cascadelakex metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/cascadelakex/clx-metrics.json | 139 +++++++++++++----- 1 file changed, 100 insertions(+), 39 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/cascadelakex/clx-metrics.json b= /tools/perf/pmu-events/arch/x86/cascadelakex/clx-metrics.json index 6485b565acbc..2e50a91b6728 100644 --- a/tools/perf/pmu-events/arch/x86/cascadelakex/clx-metrics.json +++ b/tools/perf/pmu-events/arch/x86/cascadelakex/clx-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -319,6 +319,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_0 + UOPS_DISPATCHED_PORT= .PORT_1 + UOPS_DISPATCHED_PORT.PORT_5 + UOPS_DISPATCHED_PORT.PORT_6) / tma_= info_thread_slots", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_alu_op_utilization", @@ -356,6 +357,7 @@ }, { "BriefDescription": "Total pipeline cost of instruction fetch rela= ted bottlenecks by large code footprint programs (i-side cache; TLB and BTB= misses)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * tma_fetch_latency * (tma_itlb_misses + tma_ic= ache_misses + tma_unknown_branches) / (tma_branch_resteers + tma_dsb_switch= es + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches)", "MetricGroup": "BigFootprint;BvBC;Fed;Frontend;IcMiss;MemoryTLB", "MetricName": "tma_bottleneck_big_code", @@ -369,32 +371,36 @@ "MetricThreshold": "tma_bottleneck_branching_overhead > 5", "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_4k_aliasing + tma_dtlb_load + tma_= fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + = tma_store_fwd_blk)))", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_late= ncy)) + tma_memory_bound * (tma_l3_bound / (tma_cxl_mem_bound + tma_dram_bo= und + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma= _sq_full / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_latency = + tma_sq_full)) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + t= ma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_boun= d)) * (tma_fb_full / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l= 1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_b= lk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l= 1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_b= lk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + = tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_= 4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma= _lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * = (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_boun= d + tma_store_bound)) * (tma_split_loads / (tma_4k_aliasing + tma_dtlb_load= + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_l= oads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_dra= m_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * = (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency)) + tma_memory_bound * (tma_store_bound / (tma_dram_bou= nd + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_= store_latency / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tm= a_store_latency)))", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latenc= y)) + tma_memory_bound * (tma_l3_bound / (tma_cxl_mem_bound + tma_dram_boun= d + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l= 3_hit_latency / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_lat= ency + tma_sq_full)) + tma_memory_bound * tma_l2_bound / (tma_cxl_mem_bound= + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_= bound) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + tma_dram_b= ound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tm= a_l1_latency_dependency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + = tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_= fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + tma_dra= m_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * = (tma_lock_latency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1= _latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_bl= k)) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + tma_dram_boun= d + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_s= plit_loads / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latenc= y_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + t= ma_memory_bound * (tma_store_bound / (tma_cxl_mem_bound + tma_dram_bound + = tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_split= _stores / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_stor= e_latency)) + tma_memory_bound * (tma_store_bound / (tma_cxl_mem_bound + tm= a_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound= )) * (tma_store_latency / (tma_dtlb_store + tma_false_sharing + tma_split_s= tores + tma_store_latency)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * tma_fetch_latency * (tma_ms_swit= ches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_resteer= s * (10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_misp= redicts)) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_b= ranches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + t= ma_itlb_misses + tma_lcp + tma_ms_switches)) - tma_bottleneck_big_code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", @@ -402,6 +408,7 @@ }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_clea= rs_resteers + tma_mispredicts_resteers * (10 * tma_microcode_sequencer * tm= a_other_mispredicts / tma_branch_mispredicts)) / (tma_clears_resteers + tma= _mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma= _dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_swit= ches) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_m= ispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_nukes = / tma_other_nukes + tma_core_bound * (tma_serializing_operation + tma_core_= bound * RS_EVENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_utilized_= 0) / (tma_divider + tma_ports_utilization + tma_serializing_operation) + tm= a_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequence= r) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", @@ -410,7 +417,8 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", - "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + = tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, tma_4k_aliasing + tm= a_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + = tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound= / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store= _bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_spli= t_stores + tma_store_latency)))", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bou= nd + tma_l3_bound + tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, = tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency += tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_boun= d * (tma_store_bound / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound += tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_dtlb_store / (tma_d= tlb_store + tma_false_sharing + tma_split_stores + tma_store_latency)))", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", "MetricThreshold": "tma_bottleneck_memory_data_tlbs > 20", @@ -418,7 +426,8 @@ }, { "BriefDescription": "Total pipeline cost of Memory Synchronization= related bottlenecks (data transfers and coherency updates across processor= s)", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * = (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) * tma_remote_cach= e / (tma_local_mem + tma_remote_cache + tma_remote_mem) + tma_l3_bound / (t= ma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_boun= d) * (tma_contested_accesses + tma_data_sharing) / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full) + tma_store_bound / = (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bo= und) * tma_false_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_= stores + tma_store_latency - tma_store_latency)) + tma_machine_clears * (1 = - tma_other_nukes / tma_other_nukes))", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound) * (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency= )) * tma_remote_cache / (tma_local_mem + tma_remote_cache + tma_remote_mem)= + tma_l3_bound / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_= l2_bound + tma_l3_bound + tma_store_bound) * (tma_contested_accesses + tma_= data_sharing) / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_lat= ency + tma_sq_full) + tma_store_bound / (tma_cxl_mem_bound + tma_dram_bound= + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * tma_fals= e_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_st= ore_latency - tma_store_latency)) + tma_machine_clears * (1 - tma_other_nuk= es / tma_other_nukes))", "MetricGroup": "BvMS;LockCont;Mem;Offcore;tma_issueSyncxn", "MetricName": "tma_bottleneck_memory_synchronization", "MetricThreshold": "tma_bottleneck_memory_synchronization > 10", @@ -426,6 +435,7 @@ }, { "BriefDescription": "Total pipeline cost of Branch Misprediction r= elated bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (1 - 10 * tma_microcode_sequencer * tma_other= _mispredicts / tma_branch_mispredicts) * (tma_branch_mispredicts + tma_fetc= h_latency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switc= hes + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches))", "MetricGroup": "Bad;BadSpec;BrMispredicts;BvMP;tma_issueBM", "MetricName": "tma_bottleneck_mispredictions", @@ -434,7 +444,8 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -442,6 +453,7 @@ }, { "BriefDescription": "Total pipeline cost of \"useful operations\" = - the portion of Retiring category not covered by Branching_Overhead nor Ir= regular_Overhead.", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_retiring - (BR_INST_RETIRED.ALL_BRANCHES= + 2 * BR_INST_RETIRED.NEAR_CALL + INST_RETIRED.NOP) / tma_info_thread_slot= s - tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_se= quencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "BvUW;Ret", "MetricName": "tma_bottleneck_useful_work", @@ -469,6 +481,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU retired uops originated from CISC (complex instruction set computer) in= struction", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_microcode_sequencer - tma_assists)", "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_gro= up", "MetricName": "tma_cisc", @@ -538,6 +551,15 @@ "PublicDescription": "This metric represents fraction of slots whe= re Core non-memory issues were of a bottleneck. Shortage in hardware compu= te resources; or dependencies in software's instructions are both categoriz= ed under Core Bound. Hence it may indicate the machine ran out of an out-of= -order resource; certain execution units are overloaded or dependencies in = program's data- or instruction-flow are limiting the performance (e.g. FP-c= hained long-latency arithmetic operations).", "ScaleUnit": "100%" }, + { + "BriefDescription": "This metric roughly estimates (based on idle = latencies) how often the CPU was stalled on accesses to external CXL Memory= by loads (e.g", + "MetricExpr": "(((1 - ((19 * (MEM_LOAD_L3_MISS_RETIRED.REMOTE_DRAM= * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS)) + 10 * (MEM_LO= AD_L3_MISS_RETIRED.LOCAL_DRAM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RET= IRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REMOTE_FWD * (1 + MEM_LOAD_RETIRED= .FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REMOTE_HITM = * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS))) / (19 * (MEM_L= OAD_L3_MISS_RETIRED.REMOTE_DRAM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_R= ETIRED.L1_MISS)) + 10 * (MEM_LOAD_L3_MISS_RETIRED.LOCAL_DRAM * (1 + MEM_LOA= D_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REM= OTE_FWD * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LO= AD_L3_MISS_RETIRED.REMOTE_HITM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RE= TIRED.L1_MISS)) + (25 * (MEM_LOAD_RETIRED.LOCAL_PMM * (1 + MEM_LOAD_RETIRED= .FB_HIT / MEM_LOAD_RETIRED.L1_MISS) if #has_pmem > 0 else 0) + 33 * (MEM_LO= AD_L3_MISS_RETIRED.REMOTE_PMM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RET= IRED.L1_MISS) if #has_pmem > 0 else 0))) if #has_pmem > 0 else 1)) * (CYCLE= _ACTIVITY.STALLS_L3_MISS / tma_info_thread_clks + (CYCLE_ACTIVITY.STALLS_L1= D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma_info_thread_clks - tma_l2_bou= nd) if 1e6 * (MEM_LOAD_L3_MISS_RETIRED.REMOTE_PMM + MEM_LOAD_RETIRED.LOCAL_= PMM) > MEM_LOAD_RETIRED.L1_MISS else 0) if #has_pmem > 0 else 0)", + "MetricGroup": "MemoryBound;Server;TmaL3mem;TopdownL3;tma_L3_group= ;tma_memory_bound_group", + "MetricName": "tma_cxl_mem_bound", + "MetricThreshold": "tma_cxl_mem_bound > 0.1 & (tma_memory_bound > = 0.2 & tma_backend_bound > 0.2)", + "PublicDescription": "This metric roughly estimates (based on idle= latencies) how often the CPU was stalled on accesses to external CXL Memor= y by loads (e.g. 3D-Xpoint (Crystal Ridge, a.k.a. IXP) memory, PMM - Persis= tent Memory Module [from CLX to SPR] or any other CXL Type3 Memory [EMR onw= ards]).", + "ScaleUnit": "100%" + }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to data-sharing ac= cesses", "MetricConstraint": "NO_GROUP_EVENTS", @@ -569,7 +591,7 @@ { "BriefDescription": "This metric estimates how often the CPU was s= talled on accesses to external memory (DRAM) by loads", "MetricConstraint": "NO_GROUP_EVENTS", - "MetricExpr": "CYCLE_ACTIVITY.STALLS_L3_MISS / tma_info_thread_clk= s + (CYCLE_ACTIVITY.STALLS_L1D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma_= info_thread_clks - tma_l2_bound", + "MetricExpr": "(CYCLE_ACTIVITY.STALLS_L3_MISS / tma_info_thread_cl= ks + (CYCLE_ACTIVITY.STALLS_L1D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma= _info_thread_clks - tma_l2_bound - tma_cxl_mem_bound if #has_pmem > 0 else = CYCLE_ACTIVITY.STALLS_L3_MISS / tma_info_thread_clks + (CYCLE_ACTIVITY.STAL= LS_L1D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma_info_thread_clks - tma_l= 2_bound)", "MetricGroup": "MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_me= mory_bound_group", "MetricName": "tma_dram_bound", "MetricThreshold": "tma_dram_bound > 0.1 & (tma_memory_bound > 0.2= & tma_backend_bound > 0.2)", @@ -630,7 +652,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%" }, { @@ -693,7 +715,6 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-= point (FP) vector uops fraction the CPU has retired aggregated across all v= ector widths", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umas= k\\=3D0xfc@ / UOPS_RETIRED.RETIRE_SLOTS", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_= group;tma_issue2P", "MetricName": "tma_fp_vector", @@ -768,6 +789,7 @@ }, { "BriefDescription": "Branch Misprediction Cost: Cycles representin= g fraction of TMA slots wasted per non-speculative branch misprediction (re= tired JEClear)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "tma_bottleneck_mispredictions * tma_info_thread_slo= ts / 4 / BR_MISP_RETIRED.ALL_BRANCHES / 100", "MetricGroup": "Bad;BrMispredicts;tma_issueBM", "MetricName": "tma_info_bad_spec_branch_misprediction_cost", @@ -803,6 +825,7 @@ }, { "BriefDescription": "Total pipeline cost of DSB (uop cache) hits -= subset of the Instruction_Fetch_BW Bottleneck", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_frontend_bound * (tma_fetch_bandwidth / = (tma_fetch_bandwidth + tma_fetch_latency)) * (tma_dsb / (tma_dsb + tma_mite= )))", "MetricGroup": "DSB;Fed;FetchBW;tma_issueFB", "MetricName": "tma_info_botlnk_l2_dsb_bandwidth", @@ -820,6 +843,7 @@ }, { "BriefDescription": "Total pipeline cost of Instruction Cache miss= es - subset of the Big_Code Bottleneck", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_fetch_latency * tma_icache_misses / (tma= _branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_misses += tma_lcp + tma_ms_switches))", "MetricGroup": "Fed;FetchLat;IcMiss;tma_issueFL", "MetricName": "tma_info_botlnk_l2_ic_misses", @@ -961,7 +985,6 @@ }, { "BriefDescription": "Instructions per FP Arithmetic instruction (l= ower number means higher occurrence rate)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "INST_RETIRED.ANY / (FP_ARITH_INST_RETIRED.SCALAR + = cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=3D0xfc@)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_iparith", @@ -1249,7 +1272,7 @@ "MetricName": "tma_info_memory_tlb_store_stlb_mpki" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (UOPS_EXECUTED.CORE_CYCLES_G= E_1 / 2 if #SMT_on else cpu@UOPS_EXECUTED.THREAD\\,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -1266,6 +1289,12 @@ "MetricGroup": "Fed;FetchBW", "MetricName": "tma_info_pipeline_fetch_mite" }, + { + "BriefDescription": "Average number of uops fetched from MS per cy= cle", + "MetricExpr": "IDQ.MS_UOPS / cpu@IDQ.MS_UOPS\\,cmask\\=3D1@", + "MetricGroup": "Fed;FetchLat;MicroSeq", + "MetricName": "tma_info_pipeline_fetch_ms" + }, { "BriefDescription": "Instructions per a microcode Assist invocatio= n", "MetricExpr": "INST_RETIRED.ANY / (FP_ASSIST.ANY + OTHER_ASSISTS.A= NY)", @@ -1282,7 +1311,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -1294,16 +1323,28 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, + { + "BriefDescription": "Average 3DXP Memory Bandwidth Use for reads [= GB / sec]", + "MetricExpr": "(64 * UNC_M_PMM_RPQ_INSERTS / 1e9 / tma_info_system= _time if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryBW;Server;SoC", + "MetricName": "tma_info_system_cxl_mem_read_bw" + }, + { + "BriefDescription": "Average 3DXP Memory Bandwidth Use for Writes = [GB / sec]", + "MetricExpr": "(64 * UNC_M_PMM_WPQ_INSERTS / 1e9 / tma_info_system= _time if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryBW;Server;SoC", + "MetricName": "tma_info_system_cxl_mem_write_bw" + }, { "BriefDescription": "Average external Memory Bandwidth Use for rea= ds and writes [GB / sec]", "MetricExpr": "64 * (UNC_M_CAS_COUNT.RD + UNC_M_CAS_COUNT.WR) / 1e= 9 / tma_info_system_time", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Giga Floating Point Operations Per Second", @@ -1361,6 +1402,13 @@ "MetricName": "tma_info_system_mem_parallel_reads", "PublicDescription": "Average number of parallel data read request= s to external memory. Accounts for demand loads and L1/L2 prefetches" }, + { + "BriefDescription": "Average latency of data read request to exter= nal 3D X-Point memory [in nanoseconds]", + "MetricExpr": "(1e9 * (UNC_M_PMM_RPQ_OCCUPANCY.ALL / UNC_M_PMM_RPQ= _INSERTS) / imc_0@event\\=3D0x0@ if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryLat;Server;SoC", + "MetricName": "tma_info_system_mem_pmm_read_latency", + "PublicDescription": "Average latency of data read request to exte= rnal 3D X-Point memory [in nanoseconds]. Accounts for demand loads and L1/L= 2 data-read prefetches" + }, { "BriefDescription": "Average latency of data read request to exter= nal memory (in nanoseconds)", "MetricExpr": "1e9 * (UNC_CHA_TOR_OCCUPANCY.IA_MISS_DRD / UNC_CHA_= TOR_INSERTS.IA_MISS_DRD) / (tma_info_system_socket_clks / tma_info_system_t= ime)", @@ -1499,12 +1547,13 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "min(2 * (MEM_INST_RETIRED.ALL_LOADS - MEM_LOAD_RETI= RED.FB_HIT - MEM_LOAD_RETIRED.L1_MISS) * 20 / 100, max(CYCLE_ACTIVITY.CYCLE= S_MEM_ANY - CYCLE_ACTIVITY.CYCLES_L1D_MISS, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%" }, { @@ -1541,7 +1590,7 @@ "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%" }, { @@ -1565,6 +1614,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution port for Load operations", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_2 + UOPS_DISPATCHED_PORT= .PORT_3 + UOPS_DISPATCHED_PORT.PORT_7 - UOPS_DISPATCHED_PORT.PORT_4) / (2 *= tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_load_op_utilization", @@ -1591,6 +1641,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 1 GB pages for= data load accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_load_stlb_miss * DTLB_LOAD_MISSES.WALK_COMPLETE= D_1G / (DTLB_LOAD_MISSES.WALK_COMPLETED_4K + DTLB_LOAD_MISSES.WALK_COMPLETE= D_2M_4M + DTLB_LOAD_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_load_stlb_mis= s_group", "MetricName": "tma_load_stlb_miss_1g", @@ -1599,6 +1650,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 2 or 4 MB page= s for data load accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_load_stlb_miss * DTLB_LOAD_MISSES.WALK_COMPLETE= D_2M_4M / (DTLB_LOAD_MISSES.WALK_COMPLETED_4K + DTLB_LOAD_MISSES.WALK_COMPL= ETED_2M_4M + DTLB_LOAD_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_load_stlb_mis= s_group", "MetricName": "tma_load_stlb_miss_2m", @@ -1607,6 +1659,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 4 KB pages for= data load accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_load_stlb_miss * DTLB_LOAD_MISSES.WALK_COMPLETE= D_4K / (DTLB_LOAD_MISSES.WALK_COMPLETED_4K + DTLB_LOAD_MISSES.WALK_COMPLETE= D_2M_4M + DTLB_LOAD_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_load_stlb_mis= s_group", "MetricName": "tma_load_stlb_miss_4k", @@ -1624,6 +1677,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU spent handling cache misses due to lock operations", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(12 * max(0, MEM_INST_RETIRED.LOCK_LOADS - L2_RQSTS= .ALL_RFO) + MEM_INST_RETIRED.LOCK_LOADS / MEM_INST_RETIRED.ALL_STORES * (11= * L2_RQSTS.RFO_HIT + min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTAN= DING.CYCLES_WITH_DEMAND_RFO))) / tma_info_thread_clks", "MetricGroup": "LockCont;Offcore;TopdownL4;tma_L4_group;tma_issueR= FO;tma_l1_bound_group", "MetricName": "tma_lock_latency", @@ -1648,7 +1702,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%" }, { @@ -1657,7 +1711,7 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%" }, { @@ -1681,7 +1735,6 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU was retiring uops fetched by the Microcode Sequencer (MS) unit", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / UOPS_ISSUED.ANY * IDQ.M= S_UOPS / tma_info_thread_slots", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_heavy_operatio= ns_group;tma_issueMC;tma_issueMS", "MetricName": "tma_microcode_sequencer", @@ -1691,6 +1744,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to Branch Resteers as a result of Branch Misprediction= at execution stage", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES / (BR_MISP_RETIRED.ALL= _BRANCHES + MACHINE_CLEARS.COUNT) * INT_MISC.CLEAR_RESTEER_CYCLES / tma_inf= o_thread_clks", "MetricGroup": "BadSpec;BrMispredicts;BvMP;TopdownL4;tma_L4_group;= tma_branch_resteers_group;tma_issueBM", "MetricName": "tma_mispredicts_resteers", @@ -1745,6 +1799,7 @@ }, { "BriefDescription": "This metric represents the remaining light uo= ps fraction the CPU has executed - remaining means not covered by other sib= ling nodes", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_light_operations - (tma_fp_arith + tma_m= emory_operations + tma_fused_instructions + tma_non_fused_branches))", "MetricGroup": "Pipeline;TopdownL3;tma_L3_group;tma_light_operatio= ns_group", "MetricName": "tma_other_light_ops", @@ -1754,6 +1809,7 @@ }, { "BriefDescription": "This metric estimates fraction of slots the C= PU was stalled due to other cases of misprediction (non-retired x86 branche= s or other types).", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(tma_branch_mispredicts * (1 - BR_MISP_RETIRED.A= LL_BRANCHES / (INT_MISC.CLEARS_COUNT - MACHINE_CLEARS.COUNT)), 0.0001)", "MetricGroup": "BrMispredicts;BvIO;TopdownL3;tma_L3_group;tma_bran= ch_mispredicts_group", "MetricName": "tma_other_mispredicts", @@ -1762,6 +1818,7 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU has wasted due to Nukes (Machine Clears) not related to memory ordering= .", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(tma_machine_clears * (1 - MACHINE_CLEARS.MEMORY= _ORDERING / MACHINE_CLEARS.COUNT), 0.0001)", "MetricGroup": "BvIO;Machine_Clears;TopdownL3;tma_L3_group;tma_mac= hine_clears_group", "MetricName": "tma_other_nukes", @@ -1842,6 +1899,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "((tma_ports_utilized_0 * tma_info_thread_clks + (EX= E_ACTIVITY.1_PORTS_UTIL + tma_retiring * EXE_ACTIVITY.2_PORTS_UTIL)) / tma_= info_thread_clks if ARITH.DIVIDER_ACTIVE < CYCLE_ACTIVITY.STALLS_TOTAL - CY= CLE_ACTIVITY.STALLS_MEM_ANY else (EXE_ACTIVITY.1_PORTS_UTIL + tma_retiring = * EXE_ACTIVITY.2_PORTS_UTIL) / tma_info_thread_clks)", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", @@ -1956,7 +2014,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%" }, { @@ -2013,6 +2071,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 1 GB pages for= data store accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_store_stlb_miss * DTLB_STORE_MISSES.WALK_COMPLE= TED_1G / (DTLB_STORE_MISSES.WALK_COMPLETED_4K + DTLB_STORE_MISSES.WALK_COMP= LETED_2M_4M + DTLB_STORE_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_store_stlb_mi= ss_group", "MetricName": "tma_store_stlb_miss_1g", @@ -2021,6 +2080,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 2 or 4 MB page= s for data store accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_store_stlb_miss * DTLB_STORE_MISSES.WALK_COMPLE= TED_2M_4M / (DTLB_STORE_MISSES.WALK_COMPLETED_4K + DTLB_STORE_MISSES.WALK_C= OMPLETED_2M_4M + DTLB_STORE_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_store_stlb_mi= ss_group", "MetricName": "tma_store_stlb_miss_2m", @@ -2029,6 +2089,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 4 KB pages for= data store accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_store_stlb_miss * DTLB_STORE_MISSES.WALK_COMPLE= TED_4K / (DTLB_STORE_MISSES.WALK_COMPLETED_4K + DTLB_STORE_MISSES.WALK_COMP= LETED_2M_4M + DTLB_STORE_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_store_stlb_mi= ss_group", "MetricName": "tma_store_stlb_miss_4k", --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pg1-f202.google.com (mail-pg1-f202.google.com [209.85.215.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 6C65A2749F2 for ; Mon, 18 Aug 2025 19:05:33 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.215.202 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543942; cv=none; b=eN3roYFndRqlutjrZIH0a/MA2wHAgQSOOZJJ/Funvaf49nNSr5Yw7tcJ6AU8Ch92v5sMLNsXtXqtXHUIl2nBLCKcVUTSDnkYYW6iSex/PcOEDBxng3wzG+El27SiJur2oOrw97UpQx7VH8aauCl3arBWi2EGG8xyGYPVmUa/6a0= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543942; c=relaxed/simple; bh=4+gY19tNNw4uAn6tzBcinciudK44ajQHR5ni33kEKdw=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=CPicdfMjA1UGBcRNPXRvjHraBaJc1bKe4BItyG4aVtj9bhsS+/+0Hakh/3yR9mW/G/pox9Lj9bfYvOGT0y5g2TYqZk8itVUvxNKtq11FDnbonLKInwlb29+u1RpOeTwURW/jjHz7Zf061/nKu25tw0wJWMtuL5kiIgvB6ZSOe+0= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=ko9qa/Qe; arc=none smtp.client-ip=209.85.215.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="ko9qa/Qe" Received: by mail-pg1-f202.google.com with SMTP id 41be03b00d2f7-b471758845dso3816763a12.3 for ; Mon, 18 Aug 2025 12:05:33 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543933; x=1756148733; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=eY1MRrW/9uaJ4Y1+sf9GIlRDpQ6UFHpNslCk2B5DOTE=; b=ko9qa/Qe9jnVmlcHrLJRTiMXyUewVPfp/kUPU9XINRIdYCTzAGQHaeSsOMtTZP28au 9PpLglNBr8FSzB55lIj31MDuGc7Bm8+7FXfP/t9FKBzmXOYiC1wunQurUCbmDhytbJqX CKf+V3Le/KBFxrSNowzpu4vK/zKSGjvwBhWvdu5ZE05wRZ6hKAC4APWcuJLBZwYWEeoG GS5D+LLnpVh5dt/cLcudQIcp5sl6RnuexdXl/jMdSUKPBB+D5qEgRKaPQdRvVtdzkN+i DKPtY0bsfYQquZfaSO1AnkzFH5zBBaWu02I+CNXxXYPBNzSiEUNI/uJWCJnZutakgWM5 NZSg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543933; x=1756148733; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=eY1MRrW/9uaJ4Y1+sf9GIlRDpQ6UFHpNslCk2B5DOTE=; b=WCnSz6p+ZjR0qRkPyGEWAS7/Ugy6sqQF4j417MhknrOk911R0+6W88MrbTTQhRaCTN CYXIqZDyx6LtcG0rFeDSoGPd+JXPaaYmWIhKPLw6CdbgevSsK6xJMM2BACXo+GH+LxSQ /jeKRgVL+RusJbAmGmKNNWGZduEqrydP3j6yquJpiqHRoMqmXa1rAiC9OdEqPhUDwjbj VX3BQkJViC1sWuVVbwXS2nlI/wRe2Zy21Q49X+yeaSkH+S+xPvowNGYV2otTGLKMBImL PoB3VYaDf7ukdBFL9QdV+RnY5mF/Um9fFIlINU91i31+sZlGAvji8vwYpCY8hLvGrkqA bfMw== X-Forwarded-Encrypted: i=1; AJvYcCUWyCA26yqOIzzwAKWyqGhlzu8T/xF26VT+1hSYuelSTUiUXqdCghU4DkE9NPS2iaxc09UVan4BdjclhiA=@vger.kernel.org X-Gm-Message-State: AOJu0Yzq/SY69SgFVi2nR7+zQumtbj/AZQ9cNLCWHzVQeJG7u8aUKvJC hhWws6reW9JviL/J3whvHmCuFqPMn0n/6zyvutHLEqQ4DAneYv34A5U+yJL5X6Udk/IK5tSyfCC RjHM5lr2GFw== X-Google-Smtp-Source: AGHT+IGBiUdBRUyxb623X0EOUnBqQMev37i623G08EWNY1+29oPqkdrILZ7uqU1iGnHwUUTqAwVtPe1dx+Fv X-Received: from plxm7.prod.google.com ([2002:a17:902:db07:b0:242:ff4d:cf0e]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:902:cccb:b0:23f:d903:d867 with SMTP id d9443c01a7336-2449cfab388mr6647995ad.35.1755543932684; Mon, 18 Aug 2025 12:05:32 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:02 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-7-irogers@google.com> Subject: [PATCH v2 06/20] perf vendor events: Update emeraldrapids events/metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update events from v1.14 to v1.16. Update metrics from TMA 5.0 to 5.1. The event updates come from: https://github.com/intel/perfmon/commit/9020e49e790349f86141a1f6ef6f009f189= a0f48 https://github.com/intel/perfmon/commit/a0567b56185fea4b968391aae344e9d8ed9= 782d8 Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/emeraldrapids/cache.json | 100 +++++------ .../arch/x86/emeraldrapids/emr-metrics.json | 131 ++++++++------ .../x86/emeraldrapids/floating-point.json | 43 ++--- .../arch/x86/emeraldrapids/frontend.json | 42 +++-- .../arch/x86/emeraldrapids/memory.json | 30 ++-- .../arch/x86/emeraldrapids/other.json | 28 ++- .../arch/x86/emeraldrapids/pipeline.json | 167 +++++++----------- .../arch/x86/emeraldrapids/uncore-memory.json | 82 +++++++++ .../x86/emeraldrapids/virtual-memory.json | 40 ++--- tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 10 files changed, 363 insertions(+), 302 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/emeraldrapids/cache.json b/tool= s/perf/pmu-events/arch/x86/emeraldrapids/cache.json index 10bdb193c16f..e96f938587bb 100644 --- a/tools/perf/pmu-events/arch/x86/emeraldrapids/cache.json +++ b/tools/perf/pmu-events/arch/x86/emeraldrapids/cache.json @@ -4,7 +4,6 @@ "Counter": "0,1,2,3", "EventCode": "0x51", "EventName": "L1D.HWPF_MISS", - "PublicDescription": "L1D.HWPF_MISS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -13,7 +12,7 @@ "Counter": "0,1,2,3", "EventCode": "0x51", "EventName": "L1D.REPLACEMENT", - "PublicDescription": "Counts L1D data line replacements including = opportunistic replacements, and replacements that require stall-for-replace= or block-for-replace. Available PDIST counters: 0", + "PublicDescription": "Counts L1D data line replacements including = opportunistic replacements, and replacements that require stall-for-replace= or block-for-replace.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -22,7 +21,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.FB_FULL", - "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -33,7 +32,7 @@ "EdgeDetect": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.FB_FULL_PERIODS", - "PublicDescription": "Counts number of phases a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts number of phases a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -43,7 +42,6 @@ "Deprecated": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.L2_STALL", - "PublicDescription": "This event is deprecated. Refer to new event= L1D_PEND_MISS.L2_STALLS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -52,7 +50,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.L2_STALLS", - "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D due to lack of L2 resources. Demand requests include cac= heable/uncacheable demand load, store, lock or SW prefetch accesses. Availa= ble PDIST counters: 0", + "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D due to lack of L2 resources. Demand requests include cac= heable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -61,7 +59,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.PENDING", - "PublicDescription": "Counts number of L1D misses that are outstan= ding in each cycle, that is each cycle the number of Fill Buffers (FB) outs= tanding required by Demand Reads. FB either is held by demand loads, or it = is held by non-demand loads and gets hit at least once by demand. The valid= outstanding interval is defined until the FB deallocation by one of the fo= llowing ways: from FB allocation, if FB is allocated by demand from the dem= and Hit FB, if it is allocated by hardware or software prefetch. Note: In t= he L1D, a Demand Read contains cacheable or noncacheable demand loads, incl= uding ones causing cache-line splits and reads due to page walks resulted f= rom any request type. Available PDIST counters: 0", + "PublicDescription": "Counts number of L1D misses that are outstan= ding in each cycle, that is each cycle the number of Fill Buffers (FB) outs= tanding required by Demand Reads. FB either is held by demand loads, or it = is held by non-demand loads and gets hit at least once by demand. The valid= outstanding interval is defined until the FB deallocation by one of the fo= llowing ways: from FB allocation, if FB is allocated by demand from the dem= and Hit FB, if it is allocated by hardware or software prefetch. Note: In t= he L1D, a Demand Read contains cacheable or noncacheable demand loads, incl= uding ones causing cache-line splits and reads due to page walks resulted f= rom any request type.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -71,7 +69,7 @@ "CounterMask": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.PENDING_CYCLES", - "PublicDescription": "Counts duration of L1D miss outstanding in c= ycles. Available PDIST counters: 0", + "PublicDescription": "Counts duration of L1D miss outstanding in c= ycles.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -80,7 +78,7 @@ "Counter": "0,1,2,3", "EventCode": "0x25", "EventName": "L2_LINES_IN.ALL", - "PublicDescription": "Counts the number of L2 cache lines filling = the L2. Counting does not cover rejects. Available PDIST counters: 0", + "PublicDescription": "Counts the number of L2 cache lines filling = the L2. Counting does not cover rejects.", "SampleAfterValue": "100003", "UMask": "0x1f" }, @@ -89,7 +87,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.NON_SILENT", - "PublicDescription": "Counts the number of lines that are evicted = by L2 cache when triggered by an L2 cache fill. Those lines are in Modified= state. Modified lines are written back to L3 Available PDIST counters: 0", + "PublicDescription": "Counts the number of lines that are evicted = by L2 cache when triggered by an L2 cache fill. Those lines are in Modified= state. Modified lines are written back to L3", "SampleAfterValue": "200003", "UMask": "0x2" }, @@ -98,7 +96,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.SILENT", - "PublicDescription": "Counts the number of lines that are silently= dropped by L2 cache. These lines are typically in Shared or Exclusive stat= e. A non-threaded event. Available PDIST counters: 0", + "PublicDescription": "Counts the number of lines that are silently= dropped by L2 cache. These lines are typically in Shared or Exclusive stat= e. A non-threaded event.", "SampleAfterValue": "200003", "UMask": "0x1" }, @@ -107,7 +105,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.USELESS_HWPF", - "PublicDescription": "Counts the number of cache lines that have b= een prefetched by the L2 hardware prefetcher but not used by demand access = when evicted from the L2 cache Available PDIST counters: 0", + "PublicDescription": "Counts the number of cache lines that have b= een prefetched by the L2 hardware prefetcher but not used by demand access = when evicted from the L2 cache", "SampleAfterValue": "200003", "UMask": "0x4" }, @@ -116,7 +114,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.ALL", - "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_RQSTS.REFERENCES] Available PDIST coun= ters: 0", + "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_RQSTS.REFERENCES]", "SampleAfterValue": "200003", "UMask": "0xff" }, @@ -125,7 +123,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.MISS", - "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_RQSTS.MISS] Available PDIST coun= ters: 0", + "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_RQSTS.MISS]", "SampleAfterValue": "200003", "UMask": "0x3f" }, @@ -134,7 +132,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_CODE_RD", - "PublicDescription": "Counts the total number of L2 code requests.= Available PDIST counters: 0", + "PublicDescription": "Counts the total number of L2 code requests.= ", "SampleAfterValue": "200003", "UMask": "0xe4" }, @@ -143,7 +141,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_DATA_RD", - "PublicDescription": "Counts Demand Data Read requests accessing t= he L2 cache. These requests may hit or miss L2 cache. True-miss exclude mis= ses that were merged with ongoing L2 misses. An access is counted once. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts Demand Data Read requests accessing t= he L2 cache. These requests may hit or miss L2 cache. True-miss exclude mis= ses that were merged with ongoing L2 misses. An access is counted once.", "SampleAfterValue": "200003", "UMask": "0xe1" }, @@ -152,7 +150,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_MISS", - "PublicDescription": "Counts demand requests that miss L2 cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts demand requests that miss L2 cache.", "SampleAfterValue": "200003", "UMask": "0x27" }, @@ -161,7 +159,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_REFERENCES", - "PublicDescription": "Counts demand requests to L2 cache. Availabl= e PDIST counters: 0", + "PublicDescription": "Counts demand requests to L2 cache.", "SampleAfterValue": "200003", "UMask": "0xe7" }, @@ -170,7 +168,6 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_HWPF", - "PublicDescription": "L2_RQSTS.ALL_HWPF Available PDIST counters: = 0", "SampleAfterValue": "200003", "UMask": "0xf0" }, @@ -179,7 +176,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_RFO", - "PublicDescription": "Counts the total number of RFO (read for own= ership) requests to L2 cache. L2 RFO requests include both L1D demand RFO m= isses as well as L1D RFO prefetches. Available PDIST counters: 0", + "PublicDescription": "Counts the total number of RFO (read for own= ership) requests to L2 cache. L2 RFO requests include both L1D demand RFO m= isses as well as L1D RFO prefetches.", "SampleAfterValue": "200003", "UMask": "0xe2" }, @@ -188,7 +185,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.CODE_RD_HIT", - "PublicDescription": "Counts L2 cache hits when fetching instructi= ons, code reads. Available PDIST counters: 0", + "PublicDescription": "Counts L2 cache hits when fetching instructi= ons, code reads.", "SampleAfterValue": "200003", "UMask": "0xc4" }, @@ -197,7 +194,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.CODE_RD_MISS", - "PublicDescription": "Counts L2 cache misses when fetching instruc= tions. Available PDIST counters: 0", + "PublicDescription": "Counts L2 cache misses when fetching instruc= tions.", "SampleAfterValue": "200003", "UMask": "0x24" }, @@ -206,7 +203,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT", - "PublicDescription": "Counts the number of demand Data Read reques= ts initiated by load instructions that hit L2 cache. Available PDIST counte= rs: 0", + "PublicDescription": "Counts the number of demand Data Read reques= ts initiated by load instructions that hit L2 cache.", "SampleAfterValue": "200003", "UMask": "0xc1" }, @@ -215,7 +212,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.DEMAND_DATA_RD_MISS", - "PublicDescription": "Counts demand Data Read requests with true-m= iss in the L2 cache. True-miss excludes misses that were merged with ongoin= g L2 misses. An access is counted once. Available PDIST counters: 0", + "PublicDescription": "Counts demand Data Read requests with true-m= iss in the L2 cache. True-miss excludes misses that were merged with ongoin= g L2 misses. An access is counted once.", "SampleAfterValue": "200003", "UMask": "0x21" }, @@ -224,7 +221,6 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.HWPF_MISS", - "PublicDescription": "L2_RQSTS.HWPF_MISS Available PDIST counters:= 0", "SampleAfterValue": "200003", "UMask": "0x30" }, @@ -233,7 +229,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.MISS", - "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_REQUEST.MISS] Available PDIST co= unters: 0", + "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_REQUEST.MISS]", "SampleAfterValue": "200003", "UMask": "0x3f" }, @@ -242,7 +238,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.REFERENCES", - "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_REQUEST.ALL] Available PDIST counters:= 0", + "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_REQUEST.ALL]", "SampleAfterValue": "200003", "UMask": "0xff" }, @@ -251,7 +247,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.RFO_HIT", - "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that hit L2 cache. Available PDIST counters: 0", + "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that hit L2 cache.", "SampleAfterValue": "200003", "UMask": "0xc2" }, @@ -260,7 +256,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.RFO_MISS", - "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that miss L2 cache. Available PDIST counters: 0", + "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that miss L2 cache.", "SampleAfterValue": "200003", "UMask": "0x22" }, @@ -269,7 +265,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.SWPF_HIT", - "PublicDescription": "Counts Software prefetch requests that hit t= he L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when = FB is not full. Available PDIST counters: 0", + "PublicDescription": "Counts Software prefetch requests that hit t= he L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when = FB is not full.", "SampleAfterValue": "200003", "UMask": "0xc8" }, @@ -278,7 +274,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.SWPF_MISS", - "PublicDescription": "Counts Software prefetch requests that miss = the L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when= FB is not full. Available PDIST counters: 0", + "PublicDescription": "Counts Software prefetch requests that miss = the L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when= FB is not full.", "SampleAfterValue": "200003", "UMask": "0x28" }, @@ -287,7 +283,7 @@ "Counter": "0,1,2,3", "EventCode": "0x23", "EventName": "L2_TRANS.L2_WB", - "PublicDescription": "Counts L2 writebacks that access L2 cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts L2 writebacks that access L2 cache.", "SampleAfterValue": "200003", "UMask": "0x40" }, @@ -296,7 +292,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x2e", "EventName": "LONGEST_LAT_CACHE.MISS", - "PublicDescription": "Counts core-originated cacheable requests th= at miss the L3 cache (Longest Latency cache). Requests include data and cod= e reads, Reads-for-Ownership (RFOs), speculative accesses and hardware pref= etches to the L1 and L2. It does not include hardware prefetches to the L3= , and may not count other types of requests to the L3. Available PDIST coun= ters: 0", + "PublicDescription": "Counts core-originated cacheable requests th= at miss the L3 cache (Longest Latency cache). Requests include data and cod= e reads, Reads-for-Ownership (RFOs), speculative accesses and hardware pref= etches to the L1 and L2. It does not include hardware prefetches to the L3= , and may not count other types of requests to the L3.", "SampleAfterValue": "100003", "UMask": "0x41" }, @@ -305,7 +301,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x2e", "EventName": "LONGEST_LAT_CACHE.REFERENCE", - "PublicDescription": "Counts core-originated cacheable requests to= the L3 cache (Longest Latency cache). Requests include data and code reads= , Reads-for-Ownership (RFOs), speculative accesses and hardware prefetches = to the L1 and L2. It does not include hardware prefetches to the L3, and m= ay not count other types of requests to the L3. Available PDIST counters: 0= ", + "PublicDescription": "Counts core-originated cacheable requests to= the L3 cache (Longest Latency cache). Requests include data and code reads= , Reads-for-Ownership (RFOs), speculative accesses and hardware prefetches = to the L1 and L2. It does not include hardware prefetches to the L3, and m= ay not count other types of requests to the L3.", "SampleAfterValue": "100003", "UMask": "0x4f" }, @@ -394,7 +390,7 @@ "Counter": "0,1,2,3", "EventCode": "0x43", "EventName": "MEM_LOAD_COMPLETED.L1_MISS_ANY", - "PublicDescription": "Number of completed demand load requests tha= t missed the L1 data cache including shadow misses (FB hits, merge to an on= going L1D miss) Available PDIST counters: 0", + "PublicDescription": "Number of completed demand load requests tha= t missed the L1 data cache including shadow misses (FB hits, merge to an on= going L1D miss)", "SampleAfterValue": "1000003", "UMask": "0xfd" }, @@ -563,7 +559,6 @@ "Counter": "0,1,2,3", "EventCode": "0x44", "EventName": "MEM_STORE_RETIRED.L2_HIT", - "PublicDescription": "MEM_STORE_RETIRED.L2_HIT Available PDIST cou= nters: 0", "SampleAfterValue": "200003", "UMask": "0x1" }, @@ -572,7 +567,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe5", "EventName": "MEM_UOP_RETIRED.ANY", - "PublicDescription": "Number of retired micro-operations (uops) fo= r load or store memory accesses Available PDIST counters: 0", + "PublicDescription": "Number of retired micro-operations (uops) fo= r load or store memory accesses", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -999,7 +994,6 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.ALL_REQUESTS", - "PublicDescription": "OFFCORE_REQUESTS.ALL_REQUESTS Available PDIS= T counters: 0", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -1008,7 +1002,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DATA_RD", - "PublicDescription": "Counts the demand and prefetch data reads. A= ll Core Data Reads include cacheable 'Demands' and L2 prefetchers (not L3 p= refetchers). Counting also covers reads due to page walks resulted from any= request type. Available PDIST counters: 0", + "PublicDescription": "Counts the demand and prefetch data reads. A= ll Core Data Reads include cacheable 'Demands' and L2 prefetchers (not L3 p= refetchers). Counting also covers reads due to page walks resulted from any= request type.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -1017,7 +1011,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_CODE_RD", - "PublicDescription": "Counts both cacheable and non-cacheable code= read requests. Available PDIST counters: 0", + "PublicDescription": "Counts both cacheable and non-cacheable code= read requests.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -1026,7 +1020,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_DATA_RD", - "PublicDescription": "Counts the Demand Data Read requests sent to= uncore. Use it in conjunction with OFFCORE_REQUESTS_OUTSTANDING to determi= ne average latency in the uncore. Available PDIST counters: 0", + "PublicDescription": "Counts the Demand Data Read requests sent to= uncore. Use it in conjunction with OFFCORE_REQUESTS_OUTSTANDING to determi= ne average latency in the uncore.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -1035,7 +1029,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_RFO", - "PublicDescription": "Counts the demand RFO (read for ownership) r= equests including regular RFOs, locks, ItoM. Available PDIST counters: 0", + "PublicDescription": "Counts the demand RFO (read for ownership) r= equests including regular RFOs, locks, ItoM.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -1045,7 +1039,6 @@ "Deprecated": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD", - "PublicDescription": "This event is deprecated. Refer to new event= OFFCORE_REQUESTS_OUTSTANDING.DATA_RD Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -1055,7 +1048,6 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD", - "PublicDescription": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DAT= A_RD Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -1065,7 +1057,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_CODE= _RD", - "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1075,7 +1067,6 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA= _RD", - "PublicDescription": "Cycles where at least 1 outstanding demand d= ata read request is pending. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1085,7 +1076,6 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO", - "PublicDescription": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEM= AND_RFO Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -1094,7 +1084,6 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DATA_RD", - "PublicDescription": "OFFCORE_REQUESTS_OUTSTANDING.DATA_RD Availab= le PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -1103,7 +1092,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_CODE_RD", - "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1112,7 +1101,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD", - "PublicDescription": "For every cycle, increments by the number of= outstanding demand data read requests pending. Requests are considered o= utstanding from the time they miss the core's L2 cache until the transactio= n completion message is sent to the requestor. Available PDIST counters: 0", + "PublicDescription": "For every cycle, increments by the number of= outstanding demand data read requests pending. Requests are considered o= utstanding from the time they miss the core's L2 cache until the transactio= n completion message is sent to the requestor.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -1121,7 +1110,7 @@ "Counter": "0,1,2,3", "EventCode": "0x2c", "EventName": "SQ_MISC.BUS_LOCK", - "PublicDescription": "Counts the more expensive bus lock needed to= enforce cache coherency for certain memory accesses that need to be done a= tomically. Can be created by issuing an atomic instruction (via the LOCK p= refix) which causes a cache line split or accesses uncacheable memory. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts the more expensive bus lock needed to= enforce cache coherency for certain memory accesses that need to be done a= tomically. Can be created by issuing an atomic instruction (via the LOCK p= refix) which causes a cache line split or accesses uncacheable memory.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -1130,7 +1119,6 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.ANY", - "PublicDescription": "Counts the number of PREFETCHNTA, PREFETCHW,= PREFETCHT0, PREFETCHT1 or PREFETCHT2 instructions executed. Available PDIS= T counters: 0", "SampleAfterValue": "100003", "UMask": "0xf" }, @@ -1139,7 +1127,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.NTA", - "PublicDescription": "Counts the number of PREFETCHNTA instruction= s executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHNTA instruction= s executed.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -1148,7 +1136,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.PREFETCHW", - "PublicDescription": "Counts the number of PREFETCHW instructions = executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHW instructions = executed.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -1157,7 +1145,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.T0", - "PublicDescription": "Counts the number of PREFETCHT0 instructions= executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHT0 instructions= executed.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -1166,7 +1154,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.T1_T2", - "PublicDescription": "Counts the number of PREFETCHT1 or PREFETCHT= 2 instructions executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHT1 or PREFETCHT= 2 instructions executed.", "SampleAfterValue": "100003", "UMask": "0x4" } diff --git a/tools/perf/pmu-events/arch/x86/emeraldrapids/emr-metrics.json = b/tools/perf/pmu-events/arch/x86/emeraldrapids/emr-metrics.json index 34e1cbcd722c..af0a7dd81e93 100644 --- a/tools/perf/pmu-events/arch/x86/emeraldrapids/emr-metrics.json +++ b/tools/perf/pmu-events/arch/x86/emeraldrapids/emr-metrics.json @@ -1,28 +1,28 @@ [ { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" @@ -78,6 +78,12 @@ "MetricName": "io_bandwidth_read", "ScaleUnit": "1MB/s" }, + { + "BriefDescription": "Bandwidth of inbound IO reads that are initia= ted by end device controllers that are requesting memory from the CPU and m= iss the L3 cache", + "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_MISS_PCIRDCUR * 64 / 1e6 / d= uration_time", + "MetricName": "io_bandwidth_read_l3_miss", + "ScaleUnit": "1MB/s" + }, { "BriefDescription": "Bandwidth of IO reads that are initiated by e= nd device controllers that are requesting memory from the local CPU socket", "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_PCIRDCUR_LOCAL * 64 / 1e6 / = duration_time", @@ -96,6 +102,12 @@ "MetricName": "io_bandwidth_write", "ScaleUnit": "1MB/s" }, + { + "BriefDescription": "Bandwidth of inbound IO writes that are initi= ated by end device controllers that are writing memory to the CPU", + "MetricExpr": "(UNC_CHA_TOR_INSERTS.IO_MISS_ITOM + UNC_CHA_TOR_INS= ERTS.IO_MISS_ITOMCACHENEAR) * 64 / 1e6 / duration_time", + "MetricName": "io_bandwidth_write_l3_miss", + "ScaleUnit": "1MB/s" + }, { "BriefDescription": "Bandwidth of IO writes that are initiated by = end device controllers that are writing memory to the local CPU socket", "MetricExpr": "(UNC_CHA_TOR_INSERTS.IO_ITOM_LOCAL + UNC_CHA_TOR_IN= SERTS.IO_ITOMCACHENEAR_LOCAL) * 64 / 1e6 / duration_time", @@ -111,19 +123,19 @@ { "BriefDescription": "Percentage of inbound full cacheline writes i= nitiated by end device controllers that miss the L3 cache", "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_MISS_ITOM / UNC_CHA_TOR_INSE= RTS.IO_ITOM", - "MetricName": "io_percent_of_inbound_full_writes_that_miss_l3", + "MetricName": "io_full_write_l3_miss", "ScaleUnit": "100%" }, { "BriefDescription": "Percentage of inbound partial cacheline write= s initiated by end device controllers that miss the L3 cache", "MetricExpr": "(UNC_CHA_TOR_INSERTS.IO_MISS_ITOMCACHENEAR + UNC_CH= A_TOR_INSERTS.IO_MISS_RFO) / (UNC_CHA_TOR_INSERTS.IO_ITOMCACHENEAR + UNC_CH= A_TOR_INSERTS.IO_RFO)", - "MetricName": "io_percent_of_inbound_partial_writes_that_miss_l3", + "MetricName": "io_partial_write_l3_miss", "ScaleUnit": "100%" }, { "BriefDescription": "Percentage of inbound reads initiated by end = device controllers that miss the L3 cache", "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_MISS_PCIRDCUR / UNC_CHA_TOR_= INSERTS.IO_PCIRDCUR", - "MetricName": "io_percent_of_inbound_reads_that_miss_l3", + "MetricName": "io_read_l3_miss", "ScaleUnit": "100%" }, { @@ -335,7 +347,7 @@ { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "topdown\\-be\\-bound / (topdown\\-fe\\-bound + topd= own\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_inf= o_thread_slots", + "MetricExpr": "topdown\\-be\\-bound / (topdown\\-fe\\-bound + topd= own\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BvOB;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", @@ -369,40 +381,40 @@ "MetricThreshold": "tma_bottleneck_branching_overhead > 5", "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_amx_busy= + tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_c= ore_bound * tma_amx_busy / (tma_amx_busy + tma_divider + tma_ports_utilizat= ion + tma_serializing_operation) + tma_core_bound * (tma_ports_utilization = / (tma_amx_busy + tma_divider + tma_ports_utilization + tma_serializing_ope= ration)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_utili= zed_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)= ))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + 0 / (tma_dra= m_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * t= ma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency) + tma_memory_bound= * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_b= ound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma_dat= a_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tma_l1= _bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma= _store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_laten= cy_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependen= cy + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_= bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma= _l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_dtlb_load + tma_fb= _full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tm= a_store_fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tm= a_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_split_l= oads / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_= latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_s= tore_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_sharin= g + tma_split_stores + tma_store_latency + tma_streaming_stores)) + tma_mem= ory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_boun= d + tma_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store= + tma_false_sharing + tma_split_stores + tma_store_latency + tma_streaming= _stores)))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + 0 / (tma_dram_= bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * tma= _mem_latency / (tma_mem_bandwidth + tma_mem_latency) + tma_memory_bound * (= tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound= + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses + tma_= data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * tma_= l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + t= ma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_= l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_latenc= y_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + t= ma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound = * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bo= und + tma_store_bound)) * (tma_lock_latency / (tma_dtlb_load + tma_fb_full = + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_stor= e_fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_b= ound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_split_loads /= (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_latenc= y + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_b= ound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_s= tore_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tm= a_split_stores + tma_store_latency + tma_streaming_stores)) + tma_memory_bo= und * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tm= a_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store + tma= _false_sharing + tma_split_stores + tma_store_latency + tma_streaming_store= s)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_amx_busy= + tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_c= ore_bound * tma_amx_busy / (tma_amx_busy + tma_divider + tma_ports_utilizat= ion + tma_serializing_operation) + tma_core_bound * (tma_ports_utilization = / (tma_amx_busy + tma_divider + tma_ports_utilization + tma_serializing_ope= ration)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_utili= zed_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - I= NST_RETIRED.REP_ITERATION / cpu@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetc= h_latency * (tma_ms_switches + tma_branch_resteers * (tma_clears_resteers += tma_mispredicts_resteers * tma_other_mispredicts / tma_branch_mispredicts)= / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_branches))= / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_m= isses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * tma_ms / (tma_ds= b + tma_mite + tma_ms))) - tma_bottleneck_big_code", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - I= NST_RETIRED.REP_ITERATION / cpu@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetc= h_latency * (tma_ms_switches + tma_branch_resteers * (tma_clears_resteers += tma_mispredicts_resteers * tma_other_mispredicts / tma_branch_mispredicts)= / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_branches))= / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_m= isses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottleneck_big_code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20" }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * ((1 - INST_RETIRED.REP_ITERATION / cpu@UOPS_R= ETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_switches + tma_bra= nch_resteers * (tma_clears_resteers + tma_mispredicts_resteers * tma_other_= mispredicts / tma_branch_mispredicts) / (tma_clears_resteers + tma_mispredi= cts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_dsb_swit= ches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) + t= ma_fetch_bandwidth * tma_ms / (tma_dsb + tma_mite + tma_ms)) + 10 * tma_mic= rocode_sequencer * tma_other_mispredicts / tma_branch_mispredicts * tma_bra= nch_mispredicts + tma_machine_clears * tma_other_nukes / tma_other_nukes + = tma_core_bound * (tma_serializing_operation + RS.EMPTY_RESOURCE / tma_info_= thread_clks * tma_ports_utilized_0) / (tma_amx_busy + tma_divider + tma_por= ts_utilization + tma_serializing_operation) + tma_microcode_sequencer / (tm= a_few_uops_instructions + tma_microcode_sequencer) * (tma_assists / tma_mic= rocode_sequencer) * tma_heavy_operations)", + "MetricExpr": "100 * ((1 - INST_RETIRED.REP_ITERATION / cpu@UOPS_R= ETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_switches + tma_bra= nch_resteers * (tma_clears_resteers + tma_mispredicts_resteers * tma_other_= mispredicts / tma_branch_mispredicts) / (tma_clears_resteers + tma_mispredi= cts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_dsb_swit= ches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) + t= ma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_= mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_nukes= / tma_other_nukes + tma_core_bound * (tma_serializing_operation + RS.EMPTY= _RESOURCE / tma_info_thread_clks * tma_ports_utilized_0) / (tma_amx_busy + = tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_micr= ocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) * (= tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -434,7 +446,7 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -450,7 +462,7 @@ { "BriefDescription": "This metric represents fraction of slots the = CPU has wasted due to Branch Misprediction", "DefaultMetricgroupName": "TopdownL2", - "MetricExpr": "topdown\\-br\\-mispredict / (topdown\\-fe\\-bound += topdown\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tm= a_info_thread_slots", + "MetricExpr": "topdown\\-br\\-mispredict / (topdown\\-fe\\-bound += topdown\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BadSpec;BrMispredicts;BvMP;Default;TmaL2;TopdownL2= ;tma_L2_group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_specula= tion > 0.15", @@ -551,7 +563,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to contested acces= ses", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(76.6 * tma_info_system_core_frequency * (MEM_LOAD_= L3_HIT_RETIRED.XSNP_FWD * (OCR.DEMAND_DATA_RD.L3_HIT.SNOOP_HITM / (OCR.DEMA= ND_DATA_RD.L3_HIT.SNOOP_HITM + OCR.DEMAND_DATA_RD.L3_HIT.SNOOP_HIT_WITH_FWD= ))) + 74.6 * tma_info_system_core_frequency * MEM_LOAD_L3_HIT_RETIRED.XSNP_= MISS) * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS / 2) / tma_= info_thread_clks", "MetricGroup": "BvMS;DataSharing;LockCont;Offcore;Snoop;TopdownL4;= tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", @@ -658,7 +669,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%" }, { @@ -786,7 +797,7 @@ { "BriefDescription": "This metric represents fraction of slots wher= e the CPU was retiring heavy-weight operations -- instructions that require= two or more uops or micro-coded sequences", "DefaultMetricgroupName": "TopdownL2", - "MetricExpr": "topdown\\-heavy\\-ops / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_in= fo_thread_slots", + "MetricExpr": "topdown\\-heavy\\-ops / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "Default;Retire;TmaL2;TopdownL2;tma_L2_group;tma_re= tiring_group", "MetricName": "tma_heavy_operations", "MetricThreshold": "tma_heavy_operations > 0.1", @@ -1297,19 +1308,19 @@ { "BriefDescription": "Off-core accesses per kilo instruction for mo= dified write requests", "MetricExpr": "1e3 * OCR.MODIFIED_WRITE.ANY_RESPONSE / tma_info_in= st_mix_instructions", - "MetricGroup": "Offcore", + "MetricGroup": "Offcore;Server", "MetricName": "tma_info_memory_mix_offcore_mwrite_any_pki" }, { "BriefDescription": "Off-core accesses per kilo instruction for re= ads-to-core requests (speculative; including in-core HW prefetches)", "MetricExpr": "1e3 * OCR.READS_TO_CORE.ANY_RESPONSE / tma_info_ins= t_mix_instructions", - "MetricGroup": "CacheHits;Offcore", + "MetricGroup": "CacheHits;Offcore;Server", "MetricName": "tma_info_memory_mix_offcore_read_any_pki" }, { "BriefDescription": "L3 cache misses per kilo instruction for read= s-to-core requests (speculative; including in-core HW prefetches)", "MetricExpr": "1e3 * OCR.READS_TO_CORE.L3_MISS / tma_info_inst_mix= _instructions", - "MetricGroup": "Offcore", + "MetricGroup": "Offcore;Server", "MetricName": "tma_info_memory_mix_offcore_read_l3m_pki" }, { @@ -1335,21 +1346,21 @@ { "BriefDescription": "Average DRAM BW for Reads-to-Core (R2C) cover= ing for memory attached to local- and remote-socket", "MetricExpr": "64 * OCR.READS_TO_CORE.DRAM / 1e9 / tma_info_system= _time", - "MetricGroup": "HPC;Mem;MemoryBW;SoC", + "MetricGroup": "HPC;Mem;MemoryBW;Offcore;Server", "MetricName": "tma_info_memory_soc_r2c_dram_bw", "PublicDescription": "Average DRAM BW for Reads-to-Core (R2C) cove= ring for memory attached to local- and remote-socket. See R2C_Offcore_BW." }, { "BriefDescription": "Average L3-cache miss BW for Reads-to-Core (R= 2C)", "MetricExpr": "64 * OCR.READS_TO_CORE.L3_MISS / 1e9 / tma_info_sys= tem_time", - "MetricGroup": "HPC;Mem;MemoryBW;SoC", + "MetricGroup": "HPC;Mem;MemoryBW;Offcore;Server", "MetricName": "tma_info_memory_soc_r2c_l3m_bw", "PublicDescription": "Average L3-cache miss BW for Reads-to-Core (= R2C). This covering going to DRAM or other memory off-chip memory tears. Se= e R2C_Offcore_BW." }, { "BriefDescription": "Average Off-core access BW for Reads-to-Core = (R2C)", "MetricExpr": "64 * OCR.READS_TO_CORE.ANY_RESPONSE / 1e9 / tma_inf= o_system_time", - "MetricGroup": "HPC;Mem;MemoryBW;SoC", + "MetricGroup": "HPC;Mem;MemoryBW;Offcore;Server", "MetricName": "tma_info_memory_soc_r2c_offcore_bw", "PublicDescription": "Average Off-core access BW for Reads-to-Core= (R2C). R2C account for demand or prefetch load/RFO/code access that fill d= ata into the Core caches." }, @@ -1379,7 +1390,7 @@ "MetricName": "tma_info_memory_tlb_store_stlb_mpki" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (UOPS_EXECUTED.CORE_CYCLES_G= E_1 / 2 if #SMT_on else cpu@UOPS_EXECUTED.THREAD\\,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -1426,7 +1437,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -1438,16 +1449,28 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, + { + "BriefDescription": "Average 3DXP Memory Bandwidth Use for reads [= GB / sec]", + "MetricExpr": "(64 * UNC_M_PMM_RPQ_INSERTS / 1e9 / tma_info_system= _time if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryBW;Server;SoC", + "MetricName": "tma_info_system_cxl_mem_read_bw" + }, + { + "BriefDescription": "Average 3DXP Memory Bandwidth Use for Writes = [GB / sec]", + "MetricExpr": "(64 * UNC_M_PMM_WPQ_INSERTS / 1e9 / tma_info_system= _time if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryBW;Server;SoC", + "MetricName": "tma_info_system_cxl_mem_write_bw" + }, { "BriefDescription": "Average external Memory Bandwidth Use for rea= ds and writes [GB / sec]", "MetricExpr": "64 * (UNC_M_CAS_COUNT.RD + UNC_M_CAS_COUNT.WR) / 1e= 9 / tma_info_system_time", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Giga Floating Point Operations Per Second", @@ -1513,7 +1536,6 @@ }, { "BriefDescription": "Average latency of data read request to exter= nal memory (in nanoseconds)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "1e9 * (UNC_CHA_TOR_OCCUPANCY.IA_MISS_DRD / UNC_CHA_= TOR_INSERTS.IA_MISS_DRD) / (tma_info_system_socket_clks / tma_info_system_t= ime)", "MetricGroup": "Mem;MemoryLat;SoC", "MetricName": "tma_info_system_mem_read_latency", @@ -1674,12 +1696,12 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", "MetricExpr": "min(2 * (MEM_INST_RETIRED.ALL_LOADS - MEM_LOAD_RETI= RED.FB_HIT - MEM_LOAD_RETIRED.L1_MISS) * 20 / 100, max(CYCLE_ACTIVITY.CYCLE= S_MEM_ANY - MEMORY_ACTIVITY.CYCLES_L1D_MISS, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%" }, { @@ -1693,7 +1715,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles wit= h demand load accesses that hit the L2 cache under unloaded scenarios (poss= ibly L2 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "4.4 * tma_info_system_core_frequency * MEM_LOAD_RET= IRED.L2_HIT * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS / 2) = / tma_info_thread_clks", "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_l2_bound_grou= p", "MetricName": "tma_l2_hit_latency", @@ -1712,12 +1733,11 @@ }, { "BriefDescription": "This metric estimates fraction of cycles with= demand load accesses that hit the L3 cache under unloaded scenarios (possi= bly L3 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "32.6 * tma_info_system_core_frequency * (MEM_LOAD_R= ETIRED.L3_HIT * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS / 2= )) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%" }, { @@ -1800,6 +1820,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU spent handling cache misses due to lock operations", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(16 * max(0, MEM_INST_RETIRED.LOCK_LOADS - L2_RQSTS= .ALL_RFO) + MEM_INST_RETIRED.LOCK_LOADS / MEM_INST_RETIRED.ALL_STORES * (10= * L2_RQSTS.RFO_HIT + min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTAN= DING.CYCLES_WITH_DEMAND_RFO))) / tma_info_thread_clks", "MetricGroup": "LockCont;Offcore;TopdownL4;tma_L4_group;tma_issueR= FO;tma_l1_bound_group", "MetricName": "tma_lock_latency", @@ -1832,7 +1853,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%" }, { @@ -1841,13 +1862,13 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%" }, { "BriefDescription": "This metric represents fraction of slots the = Memory subsystem within the Backend was a bottleneck", "DefaultMetricgroupName": "TopdownL2", - "MetricExpr": "topdown\\-mem\\-bound / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_in= fo_thread_slots", + "MetricExpr": "topdown\\-mem\\-bound / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "Backend;Default;TmaL2;TopdownL2;tma_L2_group;tma_b= ackend_bound_group", "MetricName": "tma_memory_bound", "MetricThreshold": "tma_memory_bound > 0.2 & tma_backend_bound > 0= .2", @@ -1857,7 +1878,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to LFENCE Instructions.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * MISC2_RETIRED.LFENCE / tma_info_thread_clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_memory_fence", @@ -1910,7 +1930,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "max(IDQ.MS_CYCLES_ANY, cpu@UOPS_RETIRED.MS\\,cmask\= \=3D1@ / (UOPS_RETIRED.SLOTS / UOPS_ISSUED.ANY)) / tma_info_core_core_clks = / 2", + "MetricExpr": "max(IDQ.MS_CYCLES_ANY, cpu@UOPS_RETIRED.MS\\,cmask\= \=3D1@ / (UOPS_RETIRED.SLOTS / UOPS_ISSUED.ANY)) / tma_info_core_core_clks = / 2.4", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -1945,6 +1965,7 @@ }, { "BriefDescription": "This metric represents the remaining light uo= ps fraction the CPU has executed - remaining means not covered by other sib= ling nodes", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_light_operations - (tma_fp_arith + tma_i= nt_operations + tma_memory_operations + tma_fused_instructions + tma_non_fu= sed_branches))", "MetricGroup": "Pipeline;TopdownL3;tma_L3_group;tma_light_operatio= ns_group", "MetricName": "tma_other_light_ops", @@ -2006,6 +2027,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "((tma_ports_utilized_0 * tma_info_thread_clks + (EX= E_ACTIVITY.1_PORTS_UTIL + tma_retiring * EXE_ACTIVITY.2_3_PORTS_UTIL)) / tm= a_info_thread_clks if ARITH.DIV_ACTIVE < CYCLE_ACTIVITY.STALLS_TOTAL - EXE_= ACTIVITY.BOUND_ON_LOADS else (EXE_ACTIVITY.1_PORTS_UTIL + tma_retiring * EX= E_ACTIVITY.2_3_PORTS_UTIL) / tma_info_thread_clks)", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", @@ -2015,6 +2037,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "(EXE_ACTIVITY.EXE_BOUND_0_PORTS + max(RS.EMPTY_RESO= URCE - RESOURCE_STALLS.SCOREBOARD, 0)) / tma_info_thread_clks * (CYCLE_ACTI= VITY.STALLS_TOTAL - EXE_ACTIVITY.BOUND_ON_LOADS) / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", @@ -2024,6 +2047,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "EXE_ACTIVITY.1_PORTS_UTIL / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", @@ -2033,7 +2057,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "EXE_ACTIVITY.2_PORTS_UTIL / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", @@ -2043,7 +2066,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 3 or more uops per cycle on all execution ports (Logical= Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "UOPS_EXECUTED.CYCLES_GE_3 / tma_info_thread_clks", "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_ut= ilization_group", "MetricName": "tma_ports_utilized_3m", @@ -2072,7 +2094,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_info_= thread_slots", + "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -2100,7 +2122,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to PAUSE Instructions", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "CPU_CLK_UNHALTED.PAUSE / tma_info_thread_clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_slow_pause", @@ -2132,7 +2153,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%" }, { diff --git a/tools/perf/pmu-events/arch/x86/emeraldrapids/floating-point.js= on b/tools/perf/pmu-events/arch/x86/emeraldrapids/floating-point.json index 8c9207750c82..bc475e163227 100644 --- a/tools/perf/pmu-events/arch/x86/emeraldrapids/floating-point.json +++ b/tools/perf/pmu-events/arch/x86/emeraldrapids/floating-point.json @@ -5,7 +5,6 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.FPDIV_ACTIVE", - "PublicDescription": "ARITH.FPDIV_ACTIVE Available PDIST counters:= 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -14,7 +13,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.FP", - "PublicDescription": "Counts all microcode Floating Point assists.= Available PDIST counters: 0", + "PublicDescription": "Counts all microcode Floating Point assists.= ", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -23,7 +22,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.SSE_AVX_MIX", - "PublicDescription": "ASSISTS.SSE_AVX_MIX Available PDIST counters= : 0", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -32,7 +30,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_0", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_0 [This event is al= ias to FP_ARITH_DISPATCHED.V0] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -41,7 +38,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_1", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_1 [This event is al= ias to FP_ARITH_DISPATCHED.V1] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -50,7 +46,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_5", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_5 [This event is al= ias to FP_ARITH_DISPATCHED.V2] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -59,7 +54,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V0", - "PublicDescription": "FP_ARITH_DISPATCHED.V0 [This event is alias = to FP_ARITH_DISPATCHED.PORT_0] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -68,7 +62,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V1", - "PublicDescription": "FP_ARITH_DISPATCHED.V1 [This event is alias = to FP_ARITH_DISPATCHED.PORT_1] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -77,7 +70,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V2", - "PublicDescription": "FP_ARITH_DISPATCHED.V2 [This event is alias = to FP_ARITH_DISPATCHED.PORT_5] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -86,7 +78,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 2 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as the= y perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR re= gister need to be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 2 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as the= y perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR re= gister need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -95,7 +87,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events. Available PDIST co= unters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -104,7 +96,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 = calculations per element. The DAZ and FTZ flags in the MXCSR register need = to be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 = calculations per element. The DAZ and FTZ flags in the MXCSR register need = to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -113,7 +105,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events. Available PDIST co= unters: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -122,7 +114,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.4_FLOPS", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision and 256-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 2 or/and 4 computation operations, one for each element. = Applies to SSE* and AVX* packed single precision floating-point and packed= double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL= DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB ins= tructions count twice as they perform 2 calculations per element. The DAZ a= nd FTZ flags in the MXCSR register need to be set when using these events. = Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision and 256-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 2 or/and 4 computation operations, one for each element. = Applies to SSE* and AVX* packed single precision floating-point and packed= double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL= DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB ins= tructions count twice as they perform 2 calculations per element. The DAZ a= nd FTZ flags in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x18" }, @@ -131,7 +123,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP14= FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calc= ulations per element. The DAZ and FTZ flags in the MXCSR register need to b= e set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP14= FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calc= ulations per element. The DAZ and FTZ flags in the MXCSR register need to b= e set when using these events.", "SampleAfterValue": "100003", "UMask": "0x40" }, @@ -140,7 +132,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 16 computation oper= ations, one for each element. Applies to SSE* and AVX* packed single preci= sion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP1= 4 FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 cal= culations per element. The DAZ and FTZ flags in the MXCSR register need to = be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 16 computation oper= ations, one for each element. Applies to SSE* and AVX* packed single preci= sion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP1= 4 FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 cal= culations per element. The DAZ and FTZ flags in the MXCSR register need to = be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -149,7 +141,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.8_FLOPS", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision and 512-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 8 computation operations, one for each element. Applies = to SSE* and AVX* packed single precision and double precision floating-poin= t instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RSQRT14= RCP RCP14 DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice= as they perform 2 calculations per element. The DAZ and FTZ flags in the M= XCSR register need to be set when using these events. Available PDIST count= ers: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision and 512-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 8 computation operations, one for each element. Applies = to SSE* and AVX* packed single precision and double precision floating-poin= t instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RSQRT14= RCP RCP14 DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice= as they perform 2 calculations per element. The DAZ and FTZ flags in the M= XCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x60" }, @@ -158,7 +150,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR", - "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision and double precision floating-point instructions retired; some = instructions will count twice as noted below. Each count represents 1 comp= utational operation. Applies to SSE* and AVX* scalar single precision float= ing-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB= . FM(N)ADD/SUB instructions count twice as they perform 2 calculations per= element. The DAZ and FTZ flags in the MXCSR register need to be set when u= sing these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision and double precision floating-point instructions retired; some = instructions will count twice as noted below. Each count represents 1 comp= utational operation. Applies to SSE* and AVX* scalar single precision float= ing-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB= . FM(N)ADD/SUB instructions count twice as they perform 2 calculations per= element. The DAZ and FTZ flags in the MXCSR register need to be set when u= sing these events.", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -167,7 +159,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational scalar doubl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar double precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions co= unt twice as they perform 2 calculations per element. The DAZ and FTZ flags= in the MXCSR register need to be set when using these events. Available PD= IST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar doubl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar double precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions co= unt twice as they perform 2 calculations per element. The DAZ and FTZ flags= in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -176,7 +168,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE", - "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar single precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instr= uctions count twice as they perform 2 calculations per element. The DAZ and= FTZ flags in the MXCSR register need to be set when using these events. Av= ailable PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar single precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instr= uctions count twice as they perform 2 calculations per element. The DAZ and= FTZ flags in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -185,7 +177,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.VECTOR", - "PublicDescription": "Number of any Vector retired FP arithmetic i= nstructions. The DAZ and FTZ flags in the MXCSR register need to be set wh= en using these events. Available PDIST counters: 0", + "PublicDescription": "Number of any Vector retired FP arithmetic i= nstructions. The DAZ and FTZ flags in the MXCSR register need to be set wh= en using these events.", "SampleAfterValue": "1000003", "UMask": "0xfc" }, @@ -194,7 +186,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.128B_PACKED_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.128B_PACKED_HALF Avai= lable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -203,7 +194,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.256B_PACKED_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.256B_PACKED_HALF Avai= lable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -212,7 +202,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.512B_PACKED_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.512B_PACKED_HALF Avai= lable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -221,7 +210,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.COMPLEX_SCALAR_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.COMPLEX_SCALAR_HALF A= vailable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -230,7 +218,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.SCALAR", - "PublicDescription": "FP_ARITH_INST_RETIRED2.SCALAR Available PDIS= T counters: 0", + "PublicDescription": "FP_ARITH_INST_RETIRED2.SCALAR", "SampleAfterValue": "100003", "UMask": "0x3" }, @@ -239,7 +227,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.SCALAR_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.SCALAR_HALF Available= PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -248,7 +235,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.VECTOR", - "PublicDescription": "FP_ARITH_INST_RETIRED2.VECTOR Available PDIS= T counters: 0", + "PublicDescription": "FP_ARITH_INST_RETIRED2.VECTOR", "SampleAfterValue": "100003", "UMask": "0x1c" } diff --git a/tools/perf/pmu-events/arch/x86/emeraldrapids/frontend.json b/t= ools/perf/pmu-events/arch/x86/emeraldrapids/frontend.json index 9fe9d62b867a..793c486ffabe 100644 --- a/tools/perf/pmu-events/arch/x86/emeraldrapids/frontend.json +++ b/tools/perf/pmu-events/arch/x86/emeraldrapids/frontend.json @@ -4,7 +4,7 @@ "Counter": "0,1,2,3", "EventCode": "0x60", "EventName": "BACLEARS.ANY", - "PublicDescription": "Number of times the front-end is resteered w= hen it finds a branch instruction in a fetch line. This is called Unknown B= ranch which occurs for the first time a branch instruction is fetched or wh= en the branch is not tracked by the BPU (Branch Prediction Unit) anymore. A= vailable PDIST counters: 0", + "PublicDescription": "Number of times the front-end is resteered w= hen it finds a branch instruction in a fetch line. This is called Unknown B= ranch which occurs for the first time a branch instruction is fetched or wh= en the branch is not tracked by the BPU (Branch Prediction Unit) anymore.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -13,7 +13,7 @@ "Counter": "0,1,2,3", "EventCode": "0x87", "EventName": "DECODE.LCP", - "PublicDescription": "Counts cycles that the Instruction Length de= coder (ILD) stalls occurred due to dynamically changing prefix length of th= e decoded instruction (by operand size prefix instruction 0x66, address siz= e prefix instruction 0x67 or REX.W for Intel64). Count is proportional to t= he number of prefixes in a 16B-line. This may result in a three-cycle penal= ty for each LCP (Length changing prefix) in a 16-byte chunk. Available PDIS= T counters: 0", + "PublicDescription": "Counts cycles that the Instruction Length de= coder (ILD) stalls occurred due to dynamically changing prefix length of th= e decoded instruction (by operand size prefix instruction 0x66, address siz= e prefix instruction 0x67 or REX.W for Intel64). Count is proportional to t= he number of prefixes in a 16B-line. This may result in a three-cycle penal= ty for each LCP (Length changing prefix) in a 16-byte chunk.", "SampleAfterValue": "500009", "UMask": "0x1" }, @@ -22,7 +22,6 @@ "Counter": "0,1,2,3", "EventCode": "0x87", "EventName": "DECODE.MS_BUSY", - "PublicDescription": "Cycles the Microcode Sequencer is busy. Avai= lable PDIST counters: 0", "SampleAfterValue": "500009", "UMask": "0x2" }, @@ -31,7 +30,7 @@ "Counter": "0,1,2,3", "EventCode": "0x61", "EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES", - "PublicDescription": "Decode Stream Buffer (DSB) is a Uop-cache th= at holds translations of previously fetched instructions that were decoded = by the legacy x86 decode pipeline (MITE). This event counts fetch penalty c= ycles when a transition occurs from DSB to MITE. Available PDIST counters: = 0", + "PublicDescription": "Decode Stream Buffer (DSB) is a Uop-cache th= at holds translations of previously fetched instructions that were decoded = by the legacy x86 decode pipeline (MITE). This event counts fetch penalty c= ycles when a transition occurs from DSB to MITE.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -249,7 +248,7 @@ "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE_DATA.STALLS", - "PublicDescription": "Counts cycles where a code line fetch is sta= lled due to an L1 instruction cache miss. The decode pipeline works at a 32= Byte granularity. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where a code line fetch is sta= lled due to an L1 instruction cache miss. The decode pipeline works at a 32= Byte granularity.", "SampleAfterValue": "500009", "UMask": "0x4" }, @@ -260,7 +259,6 @@ "EdgeDetect": "1", "EventCode": "0x80", "EventName": "ICACHE_DATA.STALL_PERIODS", - "PublicDescription": "ICACHE_DATA.STALL_PERIODS Available PDIST co= unters: 0", "SampleAfterValue": "500009", "UMask": "0x4" }, @@ -269,7 +267,7 @@ "Counter": "0,1,2,3", "EventCode": "0x83", "EventName": "ICACHE_TAG.STALLS", - "PublicDescription": "Counts cycles where a code fetch is stalled = due to L1 instruction cache tag miss. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where a code fetch is stalled = due to L1 instruction cache tag miss.", "SampleAfterValue": "200003", "UMask": "0x4" }, @@ -279,7 +277,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES_ANY", - "PublicDescription": "Counts the number of cycles uops were delive= red to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) p= ath. Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles uops were delive= red to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) p= ath.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -289,7 +287,7 @@ "CounterMask": "6", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES_OK", - "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the D= SB (Decode Stream Buffer) path. Count includes uops that may 'bypass' the I= DQ. Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the D= SB (Decode Stream Buffer) path. Count includes uops that may 'bypass' the I= DQ.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -298,7 +296,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.DSB_UOPS", - "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Availab= le PDIST counters: 0", + "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -308,7 +306,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES_ANY", - "PublicDescription": "Counts the number of cycles uops were delive= red to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipe= line) path. During these cycles uops are not being delivered from the Decod= e Stream Buffer (DSB). Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles uops were delive= red to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipe= line) path. During these cycles uops are not being delivered from the Decod= e Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -318,7 +316,7 @@ "CounterMask": "6", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES_OK", - "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the M= ITE (legacy decode pipeline) path. During these cycles uops are not being d= elivered from the Decode Stream Buffer (DSB). Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the M= ITE (legacy decode pipeline) path. During these cycles uops are not being d= elivered from the Decode Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -327,7 +325,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_UOPS", - "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the MITE path. This also means that uops are= not being delivered from the Decode Stream Buffer (DSB). Available PDIST c= ounters: 0", + "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the MITE path. This also means that uops are= not being delivered from the Decode Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -337,7 +335,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_CYCLES_ANY", - "PublicDescription": "Counts cycles during which uops are being de= livered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS= ) is busy. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts cycles during which uops are being de= livered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS= ) is busy. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.", "SampleAfterValue": "2000003", "UMask": "0x20" }, @@ -348,7 +346,7 @@ "EdgeDetect": "1", "EventCode": "0x79", "EventName": "IDQ.MS_SWITCHES", - "PublicDescription": "Number of switches from DSB (Decode Stream B= uffer) or MITE (legacy decode pipeline) to the Microcode Sequencer. Availab= le PDIST counters: 0", + "PublicDescription": "Number of switches from DSB (Decode Stream B= uffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -357,7 +355,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_UOPS", - "PublicDescription": "Counts the total number of uops delivered by= the Microcode Sequencer (MS). Available PDIST counters: 0", + "PublicDescription": "Counts the total number of uops delivered by= the Microcode Sequencer (MS).", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -366,7 +364,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CORE", - "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CORE] Available PDI= ST counters: 0", + "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -376,7 +374,7 @@ "CounterMask": "6", "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CYCLES_0_UOPS_DELIV.CORE", - "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CYCLES= _0_UOPS_DELIV.CORE] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CYCLES= _0_UOPS_DELIV.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -387,7 +385,7 @@ "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CYCLES_FE_WAS_OK", "Invert": "1", - "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_UOPS_N= OT_DELIVERED.CYCLES_FE_WAS_OK] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_UOPS_N= OT_DELIVERED.CYCLES_FE_WAS_OK]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -396,7 +394,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CORE", - "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_BUBBLES.CORE] Available PDIST counters= : 0", + "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_BUBBLES.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -406,7 +404,7 @@ "CounterMask": "6", "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE", - "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_BUBBLES.CYCLES_0_UOPS_DEL= IV.CORE] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_BUBBLES.CYCLES_0_UOPS_DEL= IV.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -417,7 +415,7 @@ "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK", "Invert": "1", - "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_BUBBLE= S.CYCLES_FE_WAS_OK] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_BUBBLE= S.CYCLES_FE_WAS_OK]", "SampleAfterValue": "1000003", "UMask": "0x1" } diff --git a/tools/perf/pmu-events/arch/x86/emeraldrapids/memory.json b/too= ls/perf/pmu-events/arch/x86/emeraldrapids/memory.json index 7c3f9b76d367..5e6c1f05c981 100644 --- a/tools/perf/pmu-events/arch/x86/emeraldrapids/memory.json +++ b/tools/perf/pmu-events/arch/x86/emeraldrapids/memory.json @@ -5,7 +5,6 @@ "CounterMask": "6", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L3_MISS", - "PublicDescription": "Execution stalls while L3 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x6" }, @@ -14,7 +13,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.MEMORY_ORDERING", - "PublicDescription": "Counts the number of Machine Clears detected= dye to memory ordering. Memory Ordering Machine Clears may apply when a me= mory read may not conform to the memory ordering rules of the x86 architect= ure Available PDIST counters: 0", + "PublicDescription": "Counts the number of Machine Clears detected= dye to memory ordering. Memory Ordering Machine Clears may apply when a me= mory read may not conform to the memory ordering rules of the x86 architect= ure", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -24,7 +23,6 @@ "CounterMask": "2", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.CYCLES_L1D_MISS", - "PublicDescription": "Cycles while L1 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -34,7 +32,6 @@ "CounterMask": "3", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L1D_MISS", - "PublicDescription": "Execution stalls while L1 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -44,7 +41,7 @@ "CounterMask": "5", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L2_MISS", - "PublicDescription": "Execution stalls while L2 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock). Available PDIST counters: 0", + "PublicDescription": "Execution stalls while L2 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock).", "SampleAfterValue": "1000003", "UMask": "0x5" }, @@ -54,7 +51,7 @@ "CounterMask": "9", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L3_MISS", - "PublicDescription": "Execution stalls while L3 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock). Available PDIST counters: 0", + "PublicDescription": "Execution stalls while L3 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock).", "SampleAfterValue": "1000003", "UMask": "0x9" }, @@ -478,7 +475,6 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.L3_MISS_DEMAND_DATA_RD", - "PublicDescription": "Counts demand data read requests that miss t= he L3 cache. Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -487,7 +483,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.L3_MISS_DEMAND_DATA_RD", - "PublicDescription": "For every cycle, increments by the number of= demand data read requests pending that are known to have missed the L3 cac= he. Note that this does not capture all elapsed cycles while requests are = outstanding - only cycles from when the requests were known by the requesti= ng core to have missed the L3 cache. Available PDIST counters: 0", + "PublicDescription": "For every cycle, increments by the number of= demand data read requests pending that are known to have missed the L3 cac= he. Note that this does not capture all elapsed cycles while requests are = outstanding - only cycles from when the requests were known by the requesti= ng core to have missed the L3 cache.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -505,7 +501,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_EVENTS", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to none of the previous 3 categories (e.g. interrupt). Availabl= e PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to none of the previous 3 categories (e.g. interrupt).", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -514,7 +510,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MEM", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to various memory events (e.g. read/write capacity and conflict= s). Available PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to various memory events (e.g. read/write capacity and conflict= s).", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -523,7 +519,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MEMTYPE", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to incompatible memory type. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to incompatible memory type.", "SampleAfterValue": "100003", "UMask": "0x40" }, @@ -532,7 +528,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_UNFRIENDLY", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to HLE-unfriendly instructions. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to HLE-unfriendly instructions.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -541,7 +537,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.COMMIT", - "PublicDescription": "Counts the number of times RTM commit succee= ded. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times RTM commit succee= ded.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -550,7 +546,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.START", - "PublicDescription": "Counts the number of times we entered an RTM= region. Does not count nested transactions. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times we entered an RTM= region. Does not count nested transactions.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -559,7 +555,7 @@ "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CAPACITY_READ", - "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional reads Available PDIST counters: 0", + "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional reads", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -568,7 +564,7 @@ "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CAPACITY_WRITE", - "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional writes. Available PDIST counters: 0", + "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional writes.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -577,7 +573,7 @@ "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CONFLICT", - "PublicDescription": "Counts the number of times a TSX line had a = cache conflict. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times a TSX line had a = cache conflict.", "SampleAfterValue": "100003", "UMask": "0x1" } diff --git a/tools/perf/pmu-events/arch/x86/emeraldrapids/other.json b/tool= s/perf/pmu-events/arch/x86/emeraldrapids/other.json index a58d65556609..21f49f609ed4 100644 --- a/tools/perf/pmu-events/arch/x86/emeraldrapids/other.json +++ b/tools/perf/pmu-events/arch/x86/emeraldrapids/other.json @@ -4,10 +4,34 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.PAGE_FAULT", - "PublicDescription": "ASSISTS.PAGE_FAULT Available PDIST counters:= 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, + { + "BriefDescription": "HW_INTERRUPTS.MASKED", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xcb", + "EventName": "HW_INTERRUPTS.MASKED", + "SampleAfterValue": "100003", + "UMask": "0x2" + }, + { + "BriefDescription": "HW_INTERRUPTS.PENDING_AND_MASKED", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xcb", + "EventName": "HW_INTERRUPTS.PENDING_AND_MASKED", + "SampleAfterValue": "100003", + "UMask": "0x4" + }, + { + "BriefDescription": "Number of hardware interrupts received by the= processor.", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xcb", + "EventName": "HW_INTERRUPTS.RECEIVED", + "PublicDescription": "Counts the number of hardware interruptions = received by the processor.", + "SampleAfterValue": "203", + "UMask": "0x1" + }, { "BriefDescription": "Counts streaming stores that have any type of= response.", "Counter": "0,1,2,3", @@ -25,7 +49,7 @@ "CounterMask": "1", "EventCode": "0x2d", "EventName": "XQ.FULL_CYCLES", - "PublicDescription": "number of cycles when the thread is active a= nd the uncore cannot take any further requests (for example prefetches, loa= ds or stores initiated by the Core that miss the L2 cache). Available PDIST= counters: 0", + "PublicDescription": "number of cycles when the thread is active a= nd the uncore cannot take any further requests (for example prefetches, loa= ds or stores initiated by the Core that miss the L2 cache).", "SampleAfterValue": "1000003", "UMask": "0x1" } diff --git a/tools/perf/pmu-events/arch/x86/emeraldrapids/pipeline.json b/t= ools/perf/pmu-events/arch/x86/emeraldrapids/pipeline.json index 48bec483b49a..1fa7957956df 100644 --- a/tools/perf/pmu-events/arch/x86/emeraldrapids/pipeline.json +++ b/tools/perf/pmu-events/arch/x86/emeraldrapids/pipeline.json @@ -6,7 +6,6 @@ "Deprecated": "1", "EventCode": "0xb0", "EventName": "ARITH.DIVIDER_ACTIVE", - "PublicDescription": "This event is deprecated. Refer to new event= ARITH.DIV_ACTIVE Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x9" }, @@ -16,7 +15,7 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.DIV_ACTIVE", - "PublicDescription": "Counts cycles when divide unit is busy execu= ting divide or square root operations. Accounts for integer and floating-po= int operations. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when divide unit is busy execu= ting divide or square root operations. Accounts for integer and floating-po= int operations.", "SampleAfterValue": "1000003", "UMask": "0x9" }, @@ -27,7 +26,6 @@ "Deprecated": "1", "EventCode": "0xb0", "EventName": "ARITH.FP_DIVIDER_ACTIVE", - "PublicDescription": "This event is deprecated. Refer to new event= ARITH.FPDIV_ACTIVE Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -37,7 +35,6 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.IDIV_ACTIVE", - "PublicDescription": "This event counts the cycles the integer div= ider is busy. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -48,7 +45,6 @@ "Deprecated": "1", "EventCode": "0xb0", "EventName": "ARITH.INT_DIVIDER_ACTIVE", - "PublicDescription": "This event is deprecated. Refer to new event= ARITH.IDIV_ACTIVE Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -57,7 +53,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.ANY", - "PublicDescription": "Counts the number of occurrences where a mic= rocode assist is invoked by hardware. Examples include AD (page Access Dirt= y), FP and AVX related assists. Available PDIST counters: 0", + "PublicDescription": "Counts the number of occurrences where a mic= rocode assist is invoked by hardware. Examples include AD (page Access Dirt= y), FP and AVX related assists.", "SampleAfterValue": "100003", "UMask": "0x1b" }, @@ -217,7 +213,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C01", - "PublicDescription": "Counts core clocks when the thread is in the= C0.1 light-weight slower wakeup time but more power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions. Availab= le PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.1 light-weight slower wakeup time but more power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -226,7 +222,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C02", - "PublicDescription": "Counts core clocks when the thread is in the= C0.2 light-weight faster wakeup time but less power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions. Availab= le PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.2 light-weight faster wakeup time but less power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions.", "SampleAfterValue": "2000003", "UMask": "0x20" }, @@ -235,7 +231,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C0_WAIT", - "PublicDescription": "Counts core clocks when the thread is in the= C0.1 or C0.2 power saving optimized states (TPAUSE or UMWAIT instructions)= or running the PAUSE instruction. Available PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.1 or C0.2 power saving optimized states (TPAUSE or UMWAIT instructions)= or running the PAUSE instruction.", "SampleAfterValue": "2000003", "UMask": "0x70" }, @@ -244,7 +240,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.DISTRIBUTED", - "PublicDescription": "This event distributes cycle counts between = active hyperthreads, i.e., those in C0. A hyperthread becomes inactive whe= n it executes the HLT or MWAIT instructions. If all other hyperthreads are= inactive (or disabled or do not exist), all counts are attributed to this = hyperthread. To obtain the full count when the Core is active, sum the coun= ts from each hyperthread. Available PDIST counters: 0", + "PublicDescription": "This event distributes cycle counts between = active hyperthreads, i.e., those in C0. A hyperthread becomes inactive whe= n it executes the HLT or MWAIT instructions. If all other hyperthreads are= inactive (or disabled or do not exist), all counts are attributed to this = hyperthread. To obtain the full count when the Core is active, sum the coun= ts from each hyperthread.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -253,7 +249,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE", - "PublicDescription": "Counts Core crystal clock cycles when curren= t thread is unhalted and the other thread is halted. Available PDIST counte= rs: 0", + "PublicDescription": "Counts Core crystal clock cycles when curren= t thread is unhalted and the other thread is halted.", "SampleAfterValue": "25003", "UMask": "0x2" }, @@ -262,7 +258,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.PAUSE", - "PublicDescription": "CPU_CLK_UNHALTED.PAUSE Available PDIST count= ers: 0", "SampleAfterValue": "2000003", "UMask": "0x40" }, @@ -273,7 +268,6 @@ "EdgeDetect": "1", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.PAUSE_INST", - "PublicDescription": "CPU_CLK_UNHALTED.PAUSE_INST Available PDIST = counters: 0", "SampleAfterValue": "2000003", "UMask": "0x40" }, @@ -282,7 +276,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.REF_DISTRIBUTED", - "PublicDescription": "This event distributes Core crystal clock cy= cle counts between active hyperthreads, i.e., those in C0 sleep-state. A hy= perthread becomes inactive when it executes the HLT or MWAIT instructions. = If one thread is active in a core, all counts are attributed to this hypert= hread. To obtain the full count when the Core is active, sum the counts fro= m each hyperthread. Available PDIST counters: 0", + "PublicDescription": "This event distributes Core crystal clock cy= cle counts between active hyperthreads, i.e., those in C0 sleep-state. A hy= perthread becomes inactive when it executes the HLT or MWAIT instructions. = If one thread is active in a core, all counts are attributed to this hypert= hread. To obtain the full count when the Core is active, sum the counts fro= m each hyperthread.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -299,7 +293,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.REF_TSC_P", - "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the four (eight when Hyperthr= eading is disabled) programmable counters available for other events. Note:= On all current platforms this event stops counting during 'throttling (TM)= ' states duty off periods the processor is 'halted'. The counter update is= done at a lower clock rate then the core clock the overflow status bit for= this counter may appear 'sticky'. After the counter has overflowed and so= ftware clears the overflow status bit and resets the counter to less than M= AX. The reset value to the counter is not clocked immediately so the overfl= ow status bit will flip 'high (1)' and generate another PMI (if enabled) af= ter which the reset value gets clocked into the counter. Therefore, softwar= e will get the interrupt, read the overflow status bit '1 for bit 34 while = the counter value is less than MAX. Software should ignore this case. Avail= able PDIST counters: 0", + "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the four (eight when Hyperthr= eading is disabled) programmable counters available for other events. Note:= On all current platforms this event stops counting during 'throttling (TM)= ' states duty off periods the processor is 'halted'. The counter update is= done at a lower clock rate then the core clock the overflow status bit for= this counter may appear 'sticky'. After the counter has overflowed and so= ftware clears the overflow status bit and resets the counter to less than M= AX. The reset value to the counter is not clocked immediately so the overfl= ow status bit will flip 'high (1)' and generate another PMI (if enabled) af= ter which the reset value gets clocked into the counter. Therefore, softwar= e will get the interrupt, read the overflow status bit '1 for bit 34 while = the counter value is less than MAX. Software should ignore this case.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -316,7 +310,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.THREAD_P", - "PublicDescription": "This is an architectural event that counts t= he number of thread cycles while the thread is not in a halt state. The thr= ead enters the halt state when it is running the HLT instruction. The core = frequency may change from time to time due to power or thermal throttling. = For this reason, this event may have a changing ratio with regards to wall = clock time. Available PDIST counters: 0", + "PublicDescription": "This is an architectural event that counts t= he number of thread cycles while the thread is not in a halt state. The thr= ead enters the halt state when it is running the HLT instruction. The core = frequency may change from time to time due to power or thermal throttling. = For this reason, this event may have a changing ratio with regards to wall = clock time.", "SampleAfterValue": "2000003" }, { @@ -325,7 +319,6 @@ "CounterMask": "8", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS", - "PublicDescription": "Cycles while L1 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -335,7 +328,6 @@ "CounterMask": "1", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS", - "PublicDescription": "Cycles while L2 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -345,7 +337,6 @@ "CounterMask": "16", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY", - "PublicDescription": "Cycles while memory subsystem has an outstan= ding load. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -355,7 +346,6 @@ "CounterMask": "12", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS", - "PublicDescription": "Execution stalls while L1 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0xc" }, @@ -365,7 +355,6 @@ "CounterMask": "5", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS", - "PublicDescription": "Execution stalls while L2 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x5" }, @@ -375,7 +364,6 @@ "CounterMask": "4", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL", - "PublicDescription": "Total execution stalls. Available PDIST coun= ters: 0", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -384,7 +372,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb7", "EventName": "EXE.AMX_BUSY", - "PublicDescription": "Counts the cycles where the AMX (Advance Mat= rix Extension) unit is busy performing an operation. Available PDIST counte= rs: 0", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -393,7 +380,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.1_PORTS_UTIL", - "PublicDescription": "Counts cycles during which a total of 1 uop = was executed on all ports and Reservation Station (RS) was not empty. Avail= able PDIST counters: 0", + "PublicDescription": "Counts cycles during which a total of 1 uop = was executed on all ports and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -402,7 +389,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.2_3_PORTS_UTIL", - "PublicDescription": "Cycles total of 2 or 3 uops are executed on = all ports and Reservation Station (RS) was not empty. Available PDIST count= ers: 0", "SampleAfterValue": "2000003", "UMask": "0xc" }, @@ -411,7 +397,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.2_PORTS_UTIL", - "PublicDescription": "Counts cycles during which a total of 2 uops= were executed on all ports and Reservation Station (RS) was not empty. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts cycles during which a total of 2 uops= were executed on all ports and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -420,7 +406,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.3_PORTS_UTIL", - "PublicDescription": "Cycles total of 3 uops are executed on all p= orts and Reservation Station (RS) was not empty. Available PDIST counters: = 0", + "PublicDescription": "Cycles total of 3 uops are executed on all p= orts and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -429,7 +415,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.4_PORTS_UTIL", - "PublicDescription": "Cycles total of 4 uops are executed on all p= orts and Reservation Station (RS) was not empty. Available PDIST counters: = 0", + "PublicDescription": "Cycles total of 4 uops are executed on all p= orts and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -439,7 +425,6 @@ "CounterMask": "5", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.BOUND_ON_LOADS", - "PublicDescription": "Execution stalls while memory subsystem has = an outstanding load. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x21" }, @@ -449,7 +434,7 @@ "CounterMask": "2", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.BOUND_ON_STORES", - "PublicDescription": "Counts cycles where the Store Buffer was ful= l and no loads caused an execution stall. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where the Store Buffer was ful= l and no loads caused an execution stall.", "SampleAfterValue": "1000003", "UMask": "0x40" }, @@ -458,7 +443,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.EXE_BOUND_0_PORTS", - "PublicDescription": "Number of cycles total of 0 uops executed on= all ports, Reservation Station (RS) was not empty, the Store Buffer (SB) w= as not full and there was no outstanding load. Available PDIST counters: 0", + "PublicDescription": "Number of cycles total of 0 uops executed on= all ports, Reservation Station (RS) was not empty, the Store Buffer (SB) w= as not full and there was no outstanding load.", "SampleAfterValue": "1000003", "UMask": "0x80" }, @@ -467,7 +452,7 @@ "Counter": "0,1,2,3", "EventCode": "0x75", "EventName": "INST_DECODED.DECODERS", - "PublicDescription": "Number of decoders utilized in a cycle when = the MITE (legacy decode pipeline) fetches instructions. Available PDIST cou= nters: 0", + "PublicDescription": "Number of decoders utilized in a cycle when = the MITE (legacy decode pipeline) fetches instructions.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -492,7 +477,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.MACRO_FUSED", - "PublicDescription": "INST_RETIRED.MACRO_FUSED Available PDIST cou= nters: 0", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -501,7 +485,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.NOP", - "PublicDescription": "Counts all retired NOP or ENDBR32/64 instruc= tions Available PDIST counters: 0", + "PublicDescription": "Counts all retired NOP or ENDBR32/64 instruc= tions", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -518,7 +502,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.REP_ITERATION", - "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent. Available PDIST counters: 0", + "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -529,7 +513,7 @@ "EdgeDetect": "1", "EventCode": "0xad", "EventName": "INT_MISC.CLEARS_COUNT", - "PublicDescription": "Counts the number of speculative clears due = to any type of branch misprediction or machine clears Available PDIST count= ers: 0", + "PublicDescription": "Counts the number of speculative clears due = to any type of branch misprediction or machine clears", "SampleAfterValue": "500009", "UMask": "0x1" }, @@ -538,7 +522,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.CLEAR_RESTEER_CYCLES", - "PublicDescription": "Cycles after recovery from a branch mispredi= ction or machine clear till the first uop is issued from the resteered path= . Available PDIST counters: 0", + "PublicDescription": "Cycles after recovery from a branch mispredi= ction or machine clear till the first uop is issued from the resteered path= .", "SampleAfterValue": "500009", "UMask": "0x80" }, @@ -547,7 +531,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.MBA_STALLS", - "PublicDescription": "INT_MISC.MBA_STALLS Available PDIST counters= : 0", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -556,7 +539,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.RECOVERY_CYCLES", - "PublicDescription": "Counts core cycles when the Resource allocat= or was stalled due to recovery from an earlier branch misprediction or mach= ine clear event. Available PDIST counters: 0", + "PublicDescription": "Counts core cycles when the Resource allocat= or was stalled due to recovery from an earlier branch misprediction or mach= ine clear event.", "SampleAfterValue": "500009", "UMask": "0x1" }, @@ -567,7 +550,6 @@ "EventName": "INT_MISC.UNKNOWN_BRANCH_CYCLES", "MSRIndex": "0x3F7", "MSRValue": "0x7", - "PublicDescription": "Bubble cycles of BAClear (Unknown Branch). A= vailable PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x40" }, @@ -576,7 +558,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.UOP_DROPPING", - "PublicDescription": "Estimated number of Top-down Microarchitectu= re Analysis slots that got dropped due to non front-end reasons Available P= DIST counters: 0", + "PublicDescription": "Estimated number of Top-down Microarchitectu= re Analysis slots that got dropped due to non front-end reasons", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -585,7 +567,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.128BIT", - "PublicDescription": "INT_VEC_RETIRED.128BIT Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0x13" }, @@ -594,7 +575,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.256BIT", - "PublicDescription": "INT_VEC_RETIRED.256BIT Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0xac" }, @@ -603,7 +583,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.ADD_128", - "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 128-bit vector instructions. Available PDIST counters: 0= ", + "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 128-bit vector instructions.", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -612,7 +592,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.ADD_256", - "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 256-bit vector instructions. Available PDIST counters: 0= ", + "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 256-bit vector instructions.", "SampleAfterValue": "1000003", "UMask": "0xc" }, @@ -621,7 +601,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.MUL_256", - "PublicDescription": "INT_VEC_RETIRED.MUL_256 Available PDIST coun= ters: 0", "SampleAfterValue": "1000003", "UMask": "0x80" }, @@ -630,7 +609,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.SHUFFLES", - "PublicDescription": "INT_VEC_RETIRED.SHUFFLES Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x40" }, @@ -639,7 +617,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.VNNI_128", - "PublicDescription": "INT_VEC_RETIRED.VNNI_128 Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -648,7 +625,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.VNNI_256", - "PublicDescription": "INT_VEC_RETIRED.VNNI_256 Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -657,7 +633,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.ADDRESS_ALIAS", - "PublicDescription": "Counts the number of times a load got blocke= d due to false dependencies in MOB due to partial compare on address. Avail= able PDIST counters: 0", + "PublicDescription": "Counts the number of times a load got blocke= d due to false dependencies in MOB due to partial compare on address.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -666,7 +642,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.NO_SR", - "PublicDescription": "Counts the number of times that split load o= perations are temporarily blocked because all resources for handling the sp= lit accesses are in use. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times that split load o= perations are temporarily blocked because all resources for handling the sp= lit accesses are in use.", "SampleAfterValue": "100003", "UMask": "0x88" }, @@ -675,7 +651,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.STORE_FORWARD", - "PublicDescription": "Counts the number of times where store forwa= rding was prevented for a load operation. The most common case is a load bl= ocked due to the address of memory access (partially) overlapping with a pr= eceding uncompleted store. Note: See the table of not supported store forwa= rds in the Optimization Guide. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times where store forwa= rding was prevented for a load operation. The most common case is a load bl= ocked due to the address of memory access (partially) overlapping with a pr= eceding uncompleted store. Note: See the table of not supported store forwa= rds in the Optimization Guide.", "SampleAfterValue": "100003", "UMask": "0x82" }, @@ -684,7 +660,7 @@ "Counter": "0,1,2,3", "EventCode": "0x4c", "EventName": "LOAD_HIT_PREFETCH.SWPF", - "PublicDescription": "Counts all software-prefetch load dispatches= that hit the fill buffer (FB) allocated for the software prefetch. It can = also be incremented by some lock instructions. So it should only be used wi= th profiling so that the locks can be excluded by ASM (Assembly File) inspe= ction of the nearby instructions. Available PDIST counters: 0", + "PublicDescription": "Counts all software-prefetch load dispatches= that hit the fill buffer (FB) allocated for the software prefetch. It can = also be incremented by some lock instructions. So it should only be used wi= th profiling so that the locks can be excluded by ASM (Assembly File) inspe= ction of the nearby instructions.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -694,7 +670,7 @@ "CounterMask": "1", "EventCode": "0xa8", "EventName": "LSD.CYCLES_ACTIVE", - "PublicDescription": "Counts the cycles when at least one uop is d= elivered by the LSD (Loop-stream detector). Available PDIST counters: 0", + "PublicDescription": "Counts the cycles when at least one uop is d= elivered by the LSD (Loop-stream detector).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -704,7 +680,7 @@ "CounterMask": "6", "EventCode": "0xa8", "EventName": "LSD.CYCLES_OK", - "PublicDescription": "Counts the cycles when optimal number of uop= s is delivered by the LSD (Loop-stream detector). Available PDIST counters:= 0", + "PublicDescription": "Counts the cycles when optimal number of uop= s is delivered by the LSD (Loop-stream detector).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -713,7 +689,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa8", "EventName": "LSD.UOPS", - "PublicDescription": "Counts the number of uops delivered to the b= ack-end by the LSD(Loop Stream Detector). Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops delivered to the b= ack-end by the LSD(Loop Stream Detector).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -724,7 +700,7 @@ "EdgeDetect": "1", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.COUNT", - "PublicDescription": "Counts the number of machine clears (nukes) = of any type. Available PDIST counters: 0", + "PublicDescription": "Counts the number of machine clears (nukes) = of any type.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -733,7 +709,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.SMC", - "PublicDescription": "Counts self-modifying code (SMC) detected, w= hich causes a machine clear. Available PDIST counters: 0", + "PublicDescription": "Counts self-modifying code (SMC) detected, w= hich causes a machine clear.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -742,7 +718,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe0", "EventName": "MISC2_RETIRED.LFENCE", - "PublicDescription": "number of LFENCE retired instructions Availa= ble PDIST counters: 0", + "PublicDescription": "number of LFENCE retired instructions", "SampleAfterValue": "400009", "UMask": "0x20" }, @@ -751,7 +727,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcc", "EventName": "MISC_RETIRED.LBR_INSERTS", - "PublicDescription": "Increments when an entry is added to the Las= t Branch Record (LBR) array (or removed from the array in case of RETURNs i= n call stack mode). The event requires LBR enable via IA32_DEBUGCTL MSR and= branch type selection via MSR_LBR_SELECT. Available PDIST counters: 0", + "PublicDescription": "Increments when an entry is added to the Las= t Branch Record (LBR) array (or removed from the array in case of RETURNs i= n call stack mode). The event requires LBR enable via IA32_DEBUGCTL MSR and= branch type selection via MSR_LBR_SELECT.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -760,7 +736,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa2", "EventName": "RESOURCE_STALLS.SB", - "PublicDescription": "Counts allocation stall cycles caused by the= store buffer (SB) being full. This counts cycles that the pipeline back-en= d blocked uop delivery from the front-end. Available PDIST counters: 0", + "PublicDescription": "Counts allocation stall cycles caused by the= store buffer (SB) being full. This counts cycles that the pipeline back-en= d blocked uop delivery from the front-end.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -769,7 +745,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa2", "EventName": "RESOURCE_STALLS.SCOREBOARD", - "PublicDescription": "Counts cycles where the pipeline is stalled = due to serializing operations. Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -778,7 +753,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa5", "EventName": "RS.EMPTY", - "PublicDescription": "Counts cycles during which the reservation s= tation (RS) is empty for this logical processor. This is usually caused whe= n the front-end pipeline runs into starvation periods (e.g. branch mispredi= ctions or i-cache misses) Available PDIST counters: 0", + "PublicDescription": "Counts cycles during which the reservation s= tation (RS) is empty for this logical processor. This is usually caused whe= n the front-end pipeline runs into starvation periods (e.g. branch mispredi= ctions or i-cache misses)", "SampleAfterValue": "1000003", "UMask": "0x7" }, @@ -790,7 +765,7 @@ "EventCode": "0xa5", "EventName": "RS.EMPTY_COUNT", "Invert": "1", - "PublicDescription": "Counts end of periods where the Reservation = Station (RS) was empty. Could be useful to closely sample on front-end late= ncy issues (see the FRONTEND_RETIRED event of designated precise events) Av= ailable PDIST counters: 0", + "PublicDescription": "Counts end of periods where the Reservation = Station (RS) was empty. Could be useful to closely sample on front-end late= ncy issues (see the FRONTEND_RETIRED event of designated precise events)", "SampleAfterValue": "100003", "UMask": "0x7" }, @@ -799,7 +774,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa5", "EventName": "RS.EMPTY_RESOURCE", - "PublicDescription": "Cycles when Reservation Station (RS) is empt= y due to a resource in the back-end Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -812,7 +786,6 @@ "EventCode": "0xa5", "EventName": "RS_EMPTY.COUNT", "Invert": "1", - "PublicDescription": "This event is deprecated. Refer to new event= RS.EMPTY_COUNT Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x7" }, @@ -822,7 +795,6 @@ "Deprecated": "1", "EventCode": "0xa5", "EventName": "RS_EMPTY.CYCLES", - "PublicDescription": "This event is deprecated. Refer to new event= RS.EMPTY Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x7" }, @@ -831,7 +803,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.BACKEND_BOUND_SLOTS", - "PublicDescription": "Number of slots in TMA method where no micro= -operations were being issued from front-end to back-end of the machine due= to lack of back-end resources. Available PDIST counters: 0", + "PublicDescription": "Number of slots in TMA method where no micro= -operations were being issued from front-end to back-end of the machine due= to lack of back-end resources.", "SampleAfterValue": "10000003", "UMask": "0x2" }, @@ -840,7 +812,7 @@ "Counter": "0", "EventCode": "0xa4", "EventName": "TOPDOWN.BAD_SPEC_SLOTS", - "PublicDescription": "Number of slots of TMA method that were wast= ed due to incorrect speculation. It covers all types of control-flow or dat= a-related mis-speculations. Available PDIST counters: 0", + "PublicDescription": "Number of slots of TMA method that were wast= ed due to incorrect speculation. It covers all types of control-flow or dat= a-related mis-speculations.", "SampleAfterValue": "10000003", "UMask": "0x4" }, @@ -849,7 +821,7 @@ "Counter": "0", "EventCode": "0xa4", "EventName": "TOPDOWN.BR_MISPREDICT_SLOTS", - "PublicDescription": "Number of TMA slots that were wasted due to = incorrect speculation by (any type of) branch mispredictions. This event es= timates number of speculative operations that were issued but not retired a= s well as the out-of-order engine recovery past a branch misprediction. Ava= ilable PDIST counters: 0", + "PublicDescription": "Number of TMA slots that were wasted due to = incorrect speculation by (any type of) branch mispredictions. This event es= timates number of speculative operations that were issued but not retired a= s well as the out-of-order engine recovery past a branch misprediction.", "SampleAfterValue": "10000003", "UMask": "0x8" }, @@ -858,7 +830,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.MEMORY_BOUND_SLOTS", - "PublicDescription": "TOPDOWN.MEMORY_BOUND_SLOTS Available PDIST c= ounters: 0", "SampleAfterValue": "10000003", "UMask": "0x10" }, @@ -875,7 +846,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.SLOTS_P", - "PublicDescription": "Counts the number of available slots for an = unhalted logical processor. The event increments by machine-width of the na= rrowest pipeline as employed by the Top-down Microarchitecture Analysis met= hod. The count is distributed among unhalted logical processors (hyper-thre= ads) who share the same physical core. Available PDIST counters: 0", + "PublicDescription": "Counts the number of available slots for an = unhalted logical processor. The event increments by machine-width of the na= rrowest pipeline as employed by the Top-down Microarchitecture Analysis met= hod. The count is distributed among unhalted logical processors (hyper-thre= ads) who share the same physical core.", "SampleAfterValue": "10000003", "UMask": "0x1" }, @@ -884,7 +855,6 @@ "Counter": "0,1,2,3", "EventCode": "0x76", "EventName": "UOPS_DECODED.DEC0_UOPS", - "PublicDescription": "UOPS_DECODED.DEC0_UOPS Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -893,7 +863,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_0", - "PublicDescription": "Number of uops dispatch to execution port 0= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 0= .", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -902,7 +872,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_1", - "PublicDescription": "Number of uops dispatch to execution port 1= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 1= .", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -911,7 +881,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_2_3_10", - "PublicDescription": "Number of uops dispatch to execution ports 2= , 3 and 10 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 2= , 3 and 10", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -920,7 +890,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_4_9", - "PublicDescription": "Number of uops dispatch to execution ports 4= and 9 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 4= and 9", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -929,7 +899,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_5_11", - "PublicDescription": "Number of uops dispatch to execution ports 5= and 11 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 5= and 11", "SampleAfterValue": "2000003", "UMask": "0x20" }, @@ -938,7 +908,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_6", - "PublicDescription": "Number of uops dispatch to execution port 6= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 6= .", "SampleAfterValue": "2000003", "UMask": "0x40" }, @@ -947,7 +917,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_7_8", - "PublicDescription": "Number of uops dispatch to execution ports = 7 and 8. Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports = 7 and 8.", "SampleAfterValue": "2000003", "UMask": "0x80" }, @@ -956,7 +926,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE", - "PublicDescription": "Counts the number of uops executed from any = thread. Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops executed from any = thread.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -966,7 +936,7 @@ "CounterMask": "1", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1", - "PublicDescription": "Counts cycles when at least 1 micro-op is ex= ecuted from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 1 micro-op is ex= ecuted from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -976,7 +946,7 @@ "CounterMask": "2", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2", - "PublicDescription": "Counts cycles when at least 2 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 2 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -986,7 +956,7 @@ "CounterMask": "3", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3", - "PublicDescription": "Counts cycles when at least 3 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 3 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -996,7 +966,7 @@ "CounterMask": "4", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4", - "PublicDescription": "Counts cycles when at least 4 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 4 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -1006,7 +976,7 @@ "CounterMask": "1", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_1", - "PublicDescription": "Cycles where at least 1 uop was executed per= -thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 1 uop was executed per= -thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1016,7 +986,7 @@ "CounterMask": "2", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_2", - "PublicDescription": "Cycles where at least 2 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 2 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1026,7 +996,7 @@ "CounterMask": "3", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_3", - "PublicDescription": "Cycles where at least 3 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 3 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1036,7 +1006,7 @@ "CounterMask": "4", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_4", - "PublicDescription": "Cycles where at least 4 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 4 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1047,7 +1017,7 @@ "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.STALLS", "Invert": "1", - "PublicDescription": "Counts cycles during which no uops were disp= atched from the Reservation Station (RS) per thread. Available PDIST counte= rs: 0", + "PublicDescription": "Counts cycles during which no uops were disp= atched from the Reservation Station (RS) per thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1059,7 +1029,6 @@ "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.STALL_CYCLES", "Invert": "1", - "PublicDescription": "This event is deprecated. Refer to new event= UOPS_EXECUTED.STALLS Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1068,7 +1037,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.THREAD", - "PublicDescription": "Counts the number of uops to be executed per= -thread each cycle. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1077,7 +1045,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.X87", - "PublicDescription": "Counts the number of x87 uops executed. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts the number of x87 uops executed.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -1086,7 +1054,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xae", "EventName": "UOPS_ISSUED.ANY", - "PublicDescription": "Counts the number of uops that the Resource = Allocation Table (RAT) issues to the Reservation Station (RS). Available PD= IST counters: 0", + "PublicDescription": "Counts the number of uops that the Resource = Allocation Table (RAT) issues to the Reservation Station (RS).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1096,7 +1064,6 @@ "CounterMask": "1", "EventCode": "0xae", "EventName": "UOPS_ISSUED.CYCLES", - "PublicDescription": "UOPS_ISSUED.CYCLES Available PDIST counters:= 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1106,7 +1073,7 @@ "CounterMask": "1", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.CYCLES", - "PublicDescription": "Counts cycles where at least one uop has ret= ired. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where at least one uop has ret= ired.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1115,7 +1082,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.HEAVY", - "PublicDescription": "Counts the number of retired micro-operation= s (uops) except the last uop of each instruction. An instruction that is de= coded into less than two uops does not contribute to the count. Available P= DIST counters: 0", + "PublicDescription": "Counts the number of retired micro-operation= s (uops) except the last uop of each instruction. An instruction that is de= coded into less than two uops does not contribute to the count.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1126,7 +1093,6 @@ "EventName": "UOPS_RETIRED.MS", "MSRIndex": "0x3F7", "MSRValue": "0x8", - "PublicDescription": "UOPS_RETIRED.MS Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -1135,7 +1101,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.SLOTS", - "PublicDescription": "Counts the retirement slots used each cycle.= Available PDIST counters: 0", + "PublicDescription": "Counts the retirement slots used each cycle.= ", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -1146,7 +1112,7 @@ "EventCode": "0xc2", "EventName": "UOPS_RETIRED.STALLS", "Invert": "1", - "PublicDescription": "This event counts cycles without actually re= tired uops. Available PDIST counters: 0", + "PublicDescription": "This event counts cycles without actually re= tired uops.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1158,7 +1124,6 @@ "EventCode": "0xc2", "EventName": "UOPS_RETIRED.STALL_CYCLES", "Invert": "1", - "PublicDescription": "This event is deprecated. Refer to new event= UOPS_RETIRED.STALLS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x2" } diff --git a/tools/perf/pmu-events/arch/x86/emeraldrapids/uncore-memory.jso= n b/tools/perf/pmu-events/arch/x86/emeraldrapids/uncore-memory.json index 68be01dad7c9..90f61c9511fc 100644 --- a/tools/perf/pmu-events/arch/x86/emeraldrapids/uncore-memory.json +++ b/tools/perf/pmu-events/arch/x86/emeraldrapids/uncore-memory.json @@ -2769,6 +2769,88 @@ "UMask": "0x3", "Unit": "iMC" }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.HIGH", + "Experimental": "1", + "PerPkg": "1", + "PublicDescription": "Number of DRAM Refreshes Issued : Counts the= number of refreshes issued.", + "UMask": "0x24", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.HIGH_ALL", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x24", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.HIGH_PCH0", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x4", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.HIGH_PCH1", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x20", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.PANIC", + "Experimental": "1", + "PerPkg": "1", + "PublicDescription": "Number of DRAM Refreshes Issued : Counts the= number of refreshes issued.", + "UMask": "0x12", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.PANIC_ALL", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x12", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.PANIC_PCH0", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x2", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.PANIC_PCH1", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x10", + "Unit": "iMC" + }, { "BriefDescription": "ECC Correctable Errors", "Counter": "0,1,2,3", diff --git a/tools/perf/pmu-events/arch/x86/emeraldrapids/virtual-memory.js= on b/tools/perf/pmu-events/arch/x86/emeraldrapids/virtual-memory.json index 3d3f88600e26..609a9549cbf3 100644 --- a/tools/perf/pmu-events/arch/x86/emeraldrapids/virtual-memory.json +++ b/tools/perf/pmu-events/arch/x86/emeraldrapids/virtual-memory.json @@ -4,7 +4,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.STLB_HIT", - "PublicDescription": "Counts loads that miss the DTLB (Data TLB) a= nd hit the STLB (Second level TLB). Available PDIST counters: 0", + "PublicDescription": "Counts loads that miss the DTLB (Data TLB) a= nd hit the STLB (Second level TLB).", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -14,7 +14,7 @@ "CounterMask": "1", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a demand load. Available PDIST cou= nters: 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a demand load.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -23,7 +23,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data loads. This implies it missed in the DTLB and furth= er levels of TLB. The page walk can end with or without a fault. Available = PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data loads. This implies it missed in the DTLB and furth= er levels of TLB. The page walk can end with or without a fault.", "SampleAfterValue": "100003", "UMask": "0xe" }, @@ -32,7 +32,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_1G", - "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -41,7 +41,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data loads. This implies address translations missed in the= DTLB and further levels of TLB. The page walk can end with or without a fa= ult. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data loads. This implies address translations missed in the= DTLB and further levels of TLB. The page walk can end with or without a fa= ult.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -50,7 +50,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -59,7 +59,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for a demand load in the PMH (Page Miss Handler) each cycle. Available PDIS= T counters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for a demand load in the PMH (Page Miss Handler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -68,7 +68,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.STLB_HIT", - "PublicDescription": "Counts stores that miss the DTLB (Data TLB) = and hit the STLB (2nd Level TLB). Available PDIST counters: 0", + "PublicDescription": "Counts stores that miss the DTLB (Data TLB) = and hit the STLB (2nd Level TLB).", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -78,7 +78,7 @@ "CounterMask": "1", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a store. Available PDIST counters:= 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a store.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -87,7 +87,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data stores. This implies it missed in the DTLB and furt= her levels of TLB. The page walk can end with or without a fault. Available= PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data stores. This implies it missed in the DTLB and furt= her levels of TLB. The page walk can end with or without a fault.", "SampleAfterValue": "100003", "UMask": "0xe" }, @@ -96,7 +96,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_1G", - "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -105,7 +105,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data stores. This implies address translations missed in th= e DTLB and further levels of TLB. The page walk can end with or without a f= ault. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data stores. This implies address translations missed in th= e DTLB and further levels of TLB. The page walk can end with or without a f= ault.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -114,7 +114,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -123,7 +123,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for a store in the PMH (Page Miss Handler) each cycle. Available PDIST coun= ters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for a store in the PMH (Page Miss Handler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -132,7 +132,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.STLB_HIT", - "PublicDescription": "Counts instruction fetch requests that miss = the ITLB (Instruction TLB) and hit the STLB (Second-level TLB). Available P= DIST counters: 0", + "PublicDescription": "Counts instruction fetch requests that miss = the ITLB (Instruction TLB) and hit the STLB (Second-level TLB).", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -142,7 +142,7 @@ "CounterMask": "1", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a code (instruction fetch) request= . Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a code (instruction fetch) request= .", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -151,7 +151,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes)= caused by a code fetch. This implies it missed in the ITLB (Instruction TL= B) and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes)= caused by a code fetch. This implies it missed in the ITLB (Instruction TL= B) and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0xe" }, @@ -160,7 +160,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M page size= s) caused by a code fetch. This implies it missed in the ITLB (Instruction = TLB) and further levels of TLB. The page walk can end with or without a fau= lt. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M page size= s) caused by a code fetch. This implies it missed in the ITLB (Instruction = TLB) and further levels of TLB. The page walk can end with or without a fau= lt.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -169,7 +169,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K page sizes) = caused by a code fetch. This implies it missed in the ITLB (Instruction TLB= ) and further levels of TLB. The page walk can end with or without a fault.= Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K page sizes) = caused by a code fetch. This implies it missed in the ITLB (Instruction TLB= ) and further levels of TLB. The page walk can end with or without a fault.= ", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -178,7 +178,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for an outstanding code (instruction fetch) request in the PMH (Page Miss H= andler) each cycle. Available PDIST counters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for an outstanding code (instruction fetch) request in the PMH (Page Miss H= andler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10" } diff --git a/tools/perf/pmu-events/arch/x86/mapfile.csv b/tools/perf/pmu-ev= ents/arch/x86/mapfile.csv index 12017f568606..e3ac2dede20a 100644 --- a/tools/perf/pmu-events/arch/x86/mapfile.csv +++ b/tools/perf/pmu-events/arch/x86/mapfile.csv @@ -9,7 +9,7 @@ GenuineIntel-6-4F,v23,broadwellx,core GenuineIntel-6-55-[56789ABCDEF],v1.25,cascadelakex,core GenuineIntel-6-DD,v1.00,clearwaterforest,core GenuineIntel-6-9[6C],v1.05,elkhartlake,core -GenuineIntel-6-CF,v1.14,emeraldrapids,core +GenuineIntel-6-CF,v1.16,emeraldrapids,core GenuineIntel-6-5[CF],v13,goldmont,core GenuineIntel-6-7A,v1.01,goldmontplus,core GenuineIntel-6-B6,v1.09,grandridge,core --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pl1-f201.google.com (mail-pl1-f201.google.com [209.85.214.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 68DE9274B3E for ; Mon, 18 Aug 2025 19:05:35 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543937; cv=none; b=ZsTOaVwv9jqKywFvqYIOjT58ANqPqYfKdZbED21WwAYk9BzxaRgP86MrjKpNFNpDqu3OUpTVNv18An555PTAIbiYyLeywSiVV4IIKHkhjWGlsFU0gpGdpatwkBqBSHBCRj4JxJvCJXCV6WqgH2ZLu0ob/xrOofarzsTEkRgTLlE= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543937; c=relaxed/simple; bh=ww9Q0rfna78fAU99esFpam0P534jDoCqqbxjRdH29Iw=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=hBRNcg/w6qLn6LutGsDOYZh2rq2nHTwEDcd9LHdzEB/KW4X09tfSF2tjjDp1EQ5q35pRDGdgJW7t4CU43aE9sD8SZvhir9CgdJIjk86XLuYafGpU/hyuKmEX5pcnpEGVSFctZJ/uXzw7lV+TVSl0O7Wea0tctEZVIf3Ph1KMfio= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=PtNZ7AZQ; arc=none smtp.client-ip=209.85.214.201 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="PtNZ7AZQ" Received: by mail-pl1-f201.google.com with SMTP id d9443c01a7336-244582c20e7so50037145ad.3 for ; Mon, 18 Aug 2025 12:05:35 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543935; x=1756148735; darn=vger.kernel.org; h=to:from:subject:message-id:references:mime-version:in-reply-to:date :from:to:cc:subject:date:message-id:reply-to; bh=+eIk1qwwNZcnpEsHTorla2tdySVvzWiJvuj6iHcoNgg=; b=PtNZ7AZQA08cQMc606UdJ7G1UrbThVmCPfItUjfmU02W42yNG6O74ejAE29d/i2ATZ dA3ezw2MaW2MJkvEzuefbapQPf/lhqruIinNKPfF30Knb4xPhAoeBKC5LnvfbEWQVENr 5lk1BSrytB6i5H2lS2C3Da7tIXuX5fb9QXe0dzsS71jOw0CfqptuT88M/h1qWg9MWZNw upLsTHJkw2rMMK/OMFL8JInSD3qFNfgB1B9rRZW0DilqE4gWiJ48jVCsqzidSLeSWnnt HVDzxHoEbo1uFtw2wQ5kUfydmSmJqB6KAj31c1BAFU7fy6agvVFSMXhIjohtlTdzI6o+ B3Og== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543935; x=1756148735; h=to:from:subject:message-id:references:mime-version:in-reply-to:date :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=+eIk1qwwNZcnpEsHTorla2tdySVvzWiJvuj6iHcoNgg=; b=sBCZBx5AQMKxX4Jab1zVqXIJXkjsVDBtUBnyF2UPGA6LHCN7MmIee3S0y/L2U7CkMa bqk5BrXIciLH10y2FzhFDb8oCgBucWf/pUjjsev88i1iCSOSXQhys5aAfeV3Dh1CqkgZ 7yy8G6TU4t6Y/IUL+hENPYeKD9kHKBkISmAF9rMc71nigRktRY/34HmAB03jeNTFJOPr EV74e2lGbftUEsWPHFE+1eiHC1s15Wd/uXldkzrxkSeX48ezOb67tVNAp2IOqK7AAbeq KIQHC5DrqK0yvFTwoCtvu17HqHg7W4sVCITOlUbUWVz2D5+92kbGUknLh9YzjFXF3FUe SKYg== X-Forwarded-Encrypted: i=1; AJvYcCV3vE/0F9WZ9KNrsYlu1AU2BZywx1QBI6p5TyHpY+FozYeklGM11pOpp/EKFyV9ndDhOzg9yjVF/NhzGok=@vger.kernel.org X-Gm-Message-State: AOJu0YySJwWn52WLuFOxcD0Fg6TWQb+igJ1uHpH4boF618sqKouiqFtX jXIMApJjYEVL/g1cAtsgdBE+Gq7BlhRBOD8RvJXABWbhjuhfknuOUrA9WPm2/C2T0HXv7gYvTcB pPhApXvbCTQ== X-Google-Smtp-Source: AGHT+IGbDCrtn11FI9qIqnMtpYaF1f2zJGkcokKTjC6Dy8WTN9il8adARJvIz66InwVw1948BwzUwWlb6YDq X-Received: from pjc7.prod.google.com ([2002:a17:90b:2f47:b0:313:274d:3007]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:902:ec81:b0:242:460f:e4a2 with SMTP id d9443c01a7336-2449cf4ebb0mr5922215ad.23.1755543934649; Mon, 18 Aug 2025 12:05:34 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:03 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-8-irogers@google.com> Subject: [PATCH v2 07/20] perf vendor events: Update grandridge metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/grandridge/grr-metrics.json | 20 +++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/grandridge/grr-metrics.json b/t= ools/perf/pmu-events/arch/x86/grandridge/grr-metrics.json index 878b1caf12de..a0d637a24c1b 100644 --- a/tools/perf/pmu-events/arch/x86/grandridge/grr-metrics.json +++ b/tools/perf/pmu-events/arch/x86/grandridge/grr-metrics.json @@ -1,56 +1,56 @@ [ { "BriefDescription": "C10 residency percent per package", - "MetricExpr": "cstate_pkg@c10\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c10\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C10_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C8 residency percent per package", - "MetricExpr": "cstate_pkg@c8\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c8\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C8_Pkg_Residency", "ScaleUnit": "100%" @@ -633,7 +633,7 @@ }, { "BriefDescription": "Average CPU Utilization", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricName": "tma_info_system_cpu_utilization" }, { @@ -645,7 +645,7 @@ }, { "BriefDescription": "Fraction of cycles spent in Kernel mode", - "MetricExpr": "cpu@CPU_CLK_UNHALTED.CORE_P@k / CPU_CLK_UNHALTED.CO= RE", + "MetricExpr": "CPU_CLK_UNHALTED.CORE_P:k / CPU_CLK_UNHALTED.CORE", "MetricGroup": "Summary", "MetricName": "tma_info_system_kernel_utilization" }, --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pj1-f74.google.com (mail-pj1-f74.google.com [209.85.216.74]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id C9228274FD3 for ; Mon, 18 Aug 2025 19:05:37 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.216.74 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543950; cv=none; b=EEitFB5w+tbOQRVoMw2gTTpkrCckfJIOxdA7LyaVMMwKMPO+aVyENmgbD8B0EdVY91uVgfEPABYLCakEhdsnfIGaLYkMeBjRyRaJmE+a8fEgPrchW+MMbA9FYWn+brsgNNf65Wdm9XBV0ow8+IGrWrT/yX7hWy2rzYGrJcycE7g= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543950; c=relaxed/simple; bh=1b8FWnh8UykMyyMcDCLfPDPxWgUy655/qxKiKxovstU=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=ajLK3HNRwNfhe3KkqQ3biD8tZaVB5paIKT4NjW4RrudK8MRGRU05zW5VjsVTz5YoOeXfsRoFnNAcCRr4XWxQjfDA/YQtSad+LfyrRD2wsKoHCe5U5PC+c4BXQln5d8MfPBTlzYu/qW6kOlsh8DrB3rCGhC4yMM2AAuIQApK/wpk= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=NYtSOe2B; arc=none smtp.client-ip=209.85.216.74 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="NYtSOe2B" Received: by mail-pj1-f74.google.com with SMTP id 98e67ed59e1d1-323266a1e87so4647835a91.0 for ; Mon, 18 Aug 2025 12:05:37 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543937; x=1756148737; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=xvgYeYun29/4w+SxwN8jaV5n7f9+hp2p6B1q4KR/WeM=; b=NYtSOe2BEQMvsUUVTC2M8UNbWoCRNOykBVmFlyECx1sggLhV69ak4VF4n+E8VKH5xq v1qwx6OkHksb6NvftrJo/nPDX+orsGNofyF2Dq7LMZhPoUmw5pNI9Y4BTYGnLPJVA4Nu g5nzo1tFinR1pfz9tzAV34MaXGYyVCUcJbSemspvD3UoxHHEo/aqDjzpKCDxc+LjTKsL 3BOdJaODBl6BBjTHwXwMKOM8odAT4iIcu8uc6ZSiwskkdv9AWpXU89MxtSXWrM0n01qo bHtRCmBccbGjV9QHVqXZEbuZvx2Zx7WH9oxIr6MuDSlG2uGc1jojlB9zZsI+E195qm7R nThA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543937; x=1756148737; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=xvgYeYun29/4w+SxwN8jaV5n7f9+hp2p6B1q4KR/WeM=; b=cu8ITEuXhwAcBNqS2ynX3U/Lcf9eiw0KtJOb3mg5ieYGLr1T2g6AquWIQRpW4V1eyI eN8qXQJc1lWuZ/aaKLN3YTzLfewkENX+otxmSF7akAR32OHf22t03tdizV2pAEiKdEvg dTJkbxr7tfr9dv81ITVzy9jDYI++jyRw5uBDk+oQH47TndVEkQWc5+U7w+8ynUO2jwwg tFv3RTRuGTS/L8A1OMDUCPFLWxr8BIRj6BmRK4+Lf4Tjq5A7iZ7i23vJ24plC1N0PrEV UgYIFOb8MVqTpVguYaN7tzMXcWf94yarBuklu6FlfU+k5mN5ZZgCUprupeGlWPwAh522 qg2Q== X-Forwarded-Encrypted: i=1; AJvYcCVluVVAMV/wsxbo4sktogxznoFHxxPi5jAgZSiYZ1EBerrIujufVNPFpvck9xjP7gJ2hXQ39SPrem0DxuA=@vger.kernel.org X-Gm-Message-State: AOJu0YzZjZmJZOqr4gUTl+0g7cm6IuftFoog+eCtaczirK2etvD/zE84 K7SPm4YRrtBF+duwpPJ0/H//jG95+SjIEPC4+KQsSaNnrD6ielvvQc4dxrrIL9XO/nXoVCgzqe4 IaVyda0tg8A== X-Google-Smtp-Source: AGHT+IEU7Xno/yOnKO6ZZ5KQNg4n2qJMVo9UOwjww5hCsxjSVaSq4m47gzOxvfRbzTHMe6qq0TzLJay/1M25 X-Received: from pjgg11.prod.google.com ([2002:a17:90b:57cb:b0:31f:f3:f8c3]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:90b:2b4d:b0:31e:c1fb:dbc6 with SMTP id 98e67ed59e1d1-3234dc644d8mr13217550a91.22.1755543937094; Mon, 18 Aug 2025 12:05:37 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:04 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-9-irogers@google.com> Subject: [PATCH v2 08/20] perf vendor events: Update graniterapids events/metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update events from v1.10 to v1.12. Update metrics from TMA 5.0 to 5.1. The event updates come from: https://github.com/intel/perfmon/commit/1684fa543fd45970759bb72dc3fc00c2ef8= 7c0e8 Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/graniterapids/cache.json | 227 +++++++++++++----- .../x86/graniterapids/floating-point.json | 43 ++-- .../arch/x86/graniterapids/frontend.json | 42 ++-- .../arch/x86/graniterapids/gnr-metrics.json | 131 +++++----- .../arch/x86/graniterapids/memory.json | 33 ++- .../arch/x86/graniterapids/other.json | 30 ++- .../arch/x86/graniterapids/pipeline.json | 167 ++++++------- .../arch/x86/graniterapids/uncore-io.json | 1 - .../arch/x86/graniterapids/uncore-memory.json | 31 --- .../x86/graniterapids/virtual-memory.json | 40 +-- tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 11 files changed, 411 insertions(+), 336 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/graniterapids/cache.json b/tool= s/perf/pmu-events/arch/x86/graniterapids/cache.json index dbdeade6fe6f..7edb73583b07 100644 --- a/tools/perf/pmu-events/arch/x86/graniterapids/cache.json +++ b/tools/perf/pmu-events/arch/x86/graniterapids/cache.json @@ -4,7 +4,6 @@ "Counter": "0,1,2,3", "EventCode": "0x51", "EventName": "L1D.HWPF_MISS", - "PublicDescription": "L1D.HWPF_MISS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -13,7 +12,7 @@ "Counter": "0,1,2,3", "EventCode": "0x51", "EventName": "L1D.REPLACEMENT", - "PublicDescription": "Counts L1D data line replacements including = opportunistic replacements, and replacements that require stall-for-replace= or block-for-replace. Available PDIST counters: 0", + "PublicDescription": "Counts L1D data line replacements including = opportunistic replacements, and replacements that require stall-for-replace= or block-for-replace.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -22,7 +21,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.FB_FULL", - "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -33,7 +32,7 @@ "EdgeDetect": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.FB_FULL_PERIODS", - "PublicDescription": "Counts number of phases a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts number of phases a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -42,7 +41,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.L2_STALLS", - "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D due to lack of L2 resources. Demand requests include cac= heable/uncacheable demand load, store, lock or SW prefetch accesses. Availa= ble PDIST counters: 0", + "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D due to lack of L2 resources. Demand requests include cac= heable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -51,7 +50,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.PENDING", - "PublicDescription": "Counts number of L1D misses that are outstan= ding in each cycle, that is each cycle the number of Fill Buffers (FB) outs= tanding required by Demand Reads. FB either is held by demand loads, or it = is held by non-demand loads and gets hit at least once by demand. The valid= outstanding interval is defined until the FB deallocation by one of the fo= llowing ways: from FB allocation, if FB is allocated by demand from the dem= and Hit FB, if it is allocated by hardware or software prefetch. Note: In t= he L1D, a Demand Read contains cacheable or noncacheable demand loads, incl= uding ones causing cache-line splits and reads due to page walks resulted f= rom any request type. Available PDIST counters: 0", + "PublicDescription": "Counts number of L1D misses that are outstan= ding in each cycle, that is each cycle the number of Fill Buffers (FB) outs= tanding required by Demand Reads. FB either is held by demand loads, or it = is held by non-demand loads and gets hit at least once by demand. The valid= outstanding interval is defined until the FB deallocation by one of the fo= llowing ways: from FB allocation, if FB is allocated by demand from the dem= and Hit FB, if it is allocated by hardware or software prefetch. Note: In t= he L1D, a Demand Read contains cacheable or noncacheable demand loads, incl= uding ones causing cache-line splits and reads due to page walks resulted f= rom any request type.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -61,7 +60,7 @@ "CounterMask": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.PENDING_CYCLES", - "PublicDescription": "Counts duration of L1D miss outstanding in c= ycles. Available PDIST counters: 0", + "PublicDescription": "Counts duration of L1D miss outstanding in c= ycles.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -70,7 +69,7 @@ "Counter": "0,1,2,3", "EventCode": "0x25", "EventName": "L2_LINES_IN.ALL", - "PublicDescription": "Counts the number of L2 cache lines filling = the L2. Counting does not cover rejects. Available PDIST counters: 0", + "PublicDescription": "Counts the number of L2 cache lines filling = the L2. Counting does not cover rejects.", "SampleAfterValue": "100003", "UMask": "0x1f" }, @@ -79,7 +78,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.NON_SILENT", - "PublicDescription": "Counts the number of lines that are evicted = by L2 cache when triggered by an L2 cache fill. Those lines are in Modified= state. Modified lines are written back to L3 Available PDIST counters: 0", + "PublicDescription": "Counts the number of lines that are evicted = by L2 cache when triggered by an L2 cache fill. Those lines are in Modified= state. Modified lines are written back to L3", "SampleAfterValue": "200003", "UMask": "0x2" }, @@ -88,7 +87,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.SILENT", - "PublicDescription": "Counts the number of lines that are silently= dropped by L2 cache. These lines are typically in Shared or Exclusive stat= e. A non-threaded event. Available PDIST counters: 0", + "PublicDescription": "Counts the number of lines that are silently= dropped by L2 cache. These lines are typically in Shared or Exclusive stat= e. A non-threaded event.", "SampleAfterValue": "200003", "UMask": "0x1" }, @@ -97,7 +96,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.USELESS_HWPF", - "PublicDescription": "Counts the number of cache lines that have b= een prefetched by the L2 hardware prefetcher but not used by demand access = when evicted from the L2 cache Available PDIST counters: 0", + "PublicDescription": "Counts the number of cache lines that have b= een prefetched by the L2 hardware prefetcher but not used by demand access = when evicted from the L2 cache", "SampleAfterValue": "200003", "UMask": "0x4" }, @@ -106,7 +105,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.ALL", - "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_RQSTS.REFERENCES] Available PDIST coun= ters: 0", + "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_RQSTS.REFERENCES]", "SampleAfterValue": "200003", "UMask": "0xff" }, @@ -115,7 +114,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.HIT", - "PublicDescription": "Counts all requests that hit L2 cache. [This= event is alias to L2_RQSTS.HIT] Available PDIST counters: 0", + "PublicDescription": "Counts all requests that hit L2 cache. [This= event is alias to L2_RQSTS.HIT]", "SampleAfterValue": "200003", "UMask": "0xdf" }, @@ -124,7 +123,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.MISS", - "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_RQSTS.MISS] Available PDIST coun= ters: 0", + "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_RQSTS.MISS]", "SampleAfterValue": "200003", "UMask": "0x3f" }, @@ -133,7 +132,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_CODE_RD", - "PublicDescription": "Counts the total number of L2 code requests.= Available PDIST counters: 0", + "PublicDescription": "Counts the total number of L2 code requests.= ", "SampleAfterValue": "200003", "UMask": "0xe4" }, @@ -142,7 +141,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_DATA_RD", - "PublicDescription": "Counts Demand Data Read requests accessing t= he L2 cache. These requests may hit or miss L2 cache. True-miss exclude mis= ses that were merged with ongoing L2 misses. An access is counted once. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts Demand Data Read requests accessing t= he L2 cache. These requests may hit or miss L2 cache. True-miss exclude mis= ses that were merged with ongoing L2 misses. An access is counted once.", "SampleAfterValue": "200003", "UMask": "0xe1" }, @@ -151,7 +150,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_MISS", - "PublicDescription": "Counts demand requests that miss L2 cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts demand requests that miss L2 cache.", "SampleAfterValue": "200003", "UMask": "0x27" }, @@ -160,7 +159,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_REFERENCES", - "PublicDescription": "Counts demand requests to L2 cache. Availabl= e PDIST counters: 0", + "PublicDescription": "Counts demand requests to L2 cache.", "SampleAfterValue": "200003", "UMask": "0xe7" }, @@ -169,7 +168,6 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_HWPF", - "PublicDescription": "L2_RQSTS.ALL_HWPF Available PDIST counters: = 0", "SampleAfterValue": "200003", "UMask": "0xf0" }, @@ -178,7 +176,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_RFO", - "PublicDescription": "Counts the total number of RFO (read for own= ership) requests to L2 cache. L2 RFO requests include both L1D demand RFO m= isses as well as L1D RFO prefetches. Available PDIST counters: 0", + "PublicDescription": "Counts the total number of RFO (read for own= ership) requests to L2 cache. L2 RFO requests include both L1D demand RFO m= isses as well as L1D RFO prefetches.", "SampleAfterValue": "200003", "UMask": "0xe2" }, @@ -187,7 +185,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.CODE_RD_HIT", - "PublicDescription": "Counts L2 cache hits when fetching instructi= ons, code reads. Available PDIST counters: 0", + "PublicDescription": "Counts L2 cache hits when fetching instructi= ons, code reads.", "SampleAfterValue": "200003", "UMask": "0xc4" }, @@ -196,7 +194,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.CODE_RD_MISS", - "PublicDescription": "Counts L2 cache misses when fetching instruc= tions. Available PDIST counters: 0", + "PublicDescription": "Counts L2 cache misses when fetching instruc= tions.", "SampleAfterValue": "200003", "UMask": "0x24" }, @@ -205,7 +203,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT", - "PublicDescription": "Counts the number of demand Data Read reques= ts initiated by load instructions that hit L2 cache. Available PDIST counte= rs: 0", + "PublicDescription": "Counts the number of demand Data Read reques= ts initiated by load instructions that hit L2 cache.", "SampleAfterValue": "200003", "UMask": "0xc1" }, @@ -214,7 +212,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.DEMAND_DATA_RD_MISS", - "PublicDescription": "Counts demand Data Read requests with true-m= iss in the L2 cache. True-miss excludes misses that were merged with ongoin= g L2 misses. An access is counted once. Available PDIST counters: 0", + "PublicDescription": "Counts demand Data Read requests with true-m= iss in the L2 cache. True-miss excludes misses that were merged with ongoin= g L2 misses. An access is counted once.", "SampleAfterValue": "200003", "UMask": "0x21" }, @@ -223,7 +221,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.HIT", - "PublicDescription": "Counts all requests that hit L2 cache. [This= event is alias to L2_REQUEST.HIT] Available PDIST counters: 0", + "PublicDescription": "Counts all requests that hit L2 cache. [This= event is alias to L2_REQUEST.HIT]", "SampleAfterValue": "200003", "UMask": "0xdf" }, @@ -232,7 +230,6 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.HWPF_MISS", - "PublicDescription": "L2_RQSTS.HWPF_MISS Available PDIST counters:= 0", "SampleAfterValue": "200003", "UMask": "0x30" }, @@ -241,7 +238,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.MISS", - "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_REQUEST.MISS] Available PDIST co= unters: 0", + "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_REQUEST.MISS]", "SampleAfterValue": "200003", "UMask": "0x3f" }, @@ -250,7 +247,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.REFERENCES", - "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_REQUEST.ALL] Available PDIST counters:= 0", + "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_REQUEST.ALL]", "SampleAfterValue": "200003", "UMask": "0xff" }, @@ -259,7 +256,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.RFO_HIT", - "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that hit L2 cache. Available PDIST counters: 0", + "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that hit L2 cache.", "SampleAfterValue": "200003", "UMask": "0xc2" }, @@ -268,7 +265,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.RFO_MISS", - "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that miss L2 cache. Available PDIST counters: 0", + "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that miss L2 cache.", "SampleAfterValue": "200003", "UMask": "0x22" }, @@ -277,7 +274,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.SWPF_HIT", - "PublicDescription": "Counts Software prefetch requests that hit t= he L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when = FB is not full. Available PDIST counters: 0", + "PublicDescription": "Counts Software prefetch requests that hit t= he L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when = FB is not full.", "SampleAfterValue": "200003", "UMask": "0xc8" }, @@ -286,7 +283,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.SWPF_MISS", - "PublicDescription": "Counts Software prefetch requests that miss = the L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when= FB is not full. Available PDIST counters: 0", + "PublicDescription": "Counts Software prefetch requests that miss = the L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when= FB is not full.", "SampleAfterValue": "200003", "UMask": "0x28" }, @@ -295,7 +292,7 @@ "Counter": "0,1,2,3", "EventCode": "0x23", "EventName": "L2_TRANS.L2_WB", - "PublicDescription": "Counts L2 writebacks that access L2 cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts L2 writebacks that access L2 cache.", "SampleAfterValue": "200003", "UMask": "0x40" }, @@ -304,7 +301,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x2e", "EventName": "LONGEST_LAT_CACHE.MISS", - "PublicDescription": "Counts core-originated cacheable requests th= at miss the L3 cache (Longest Latency cache). Requests include data and cod= e reads, Reads-for-Ownership (RFOs), speculative accesses and hardware pref= etches to the L1 and L2. It does not include hardware prefetches to the L3= , and may not count other types of requests to the L3. Available PDIST coun= ters: 0", + "PublicDescription": "Counts core-originated cacheable requests th= at miss the L3 cache (Longest Latency cache). Requests include data and cod= e reads, Reads-for-Ownership (RFOs), speculative accesses and hardware pref= etches to the L1 and L2. It does not include hardware prefetches to the L3= , and may not count other types of requests to the L3.", "SampleAfterValue": "100003", "UMask": "0x41" }, @@ -313,7 +310,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x2e", "EventName": "LONGEST_LAT_CACHE.REFERENCE", - "PublicDescription": "Counts core-originated cacheable requests to= the L3 cache (Longest Latency cache). Requests include data and code reads= , Reads-for-Ownership (RFOs), speculative accesses and hardware prefetches = to the L1 and L2. It does not include hardware prefetches to the L3, and m= ay not count other types of requests to the L3. Available PDIST counters: 0= ", + "PublicDescription": "Counts core-originated cacheable requests to= the L3 cache (Longest Latency cache). Requests include data and code reads= , Reads-for-Ownership (RFOs), speculative accesses and hardware prefetches = to the L1 and L2. It does not include hardware prefetches to the L3, and m= ay not count other types of requests to the L3.", "SampleAfterValue": "100003", "UMask": "0x4f" }, @@ -437,7 +434,7 @@ "Counter": "0,1,2,3", "EventCode": "0x43", "EventName": "MEM_LOAD_COMPLETED.L1_MISS_ANY", - "PublicDescription": "Number of completed demand load requests tha= t missed the L1 data cache including shadow misses (FB hits, merge to an on= going L1D miss) Available PDIST counters: 0", + "PublicDescription": "Number of completed demand load requests tha= t missed the L1 data cache including shadow misses (FB hits, merge to an on= going L1D miss)", "SampleAfterValue": "1000003", "UMask": "0xfd" }, @@ -503,6 +500,15 @@ "SampleAfterValue": "100007", "UMask": "0x1" }, + { + "BriefDescription": "Retired load instructions with remote cxl mem= as the data source where the data request missed all caches.", + "Counter": "0,1,2,3", + "EventCode": "0xd3", + "EventName": "MEM_LOAD_L3_MISS_RETIRED.REMOTE_CXL_MEM", + "PublicDescription": "Counts retired load instructions with remote= cxl mem as the data source and the data request missed L3. Available PDIST= counters: 0", + "SampleAfterValue": "100007", + "UMask": "0x10" + }, { "BriefDescription": "MEM_LOAD_L3_MISS_RETIRED.REMOTE_DRAM", "Counter": "0,1,2,3", @@ -628,12 +634,21 @@ "SampleAfterValue": "50021", "UMask": "0x20" }, + { + "BriefDescription": "Retired load instructions with local cxl mem = as the data source where the data request missed all caches.", + "Counter": "0,1,2,3", + "Data_LA": "1", + "EventCode": "0xd1", + "EventName": "MEM_LOAD_RETIRED.LOCAL_CXL_MEM", + "PublicDescription": "Counts retired load instructions with local = cxl mem as the data source and the data request missed L3. Available PDIST = counters: 0", + "SampleAfterValue": "1000003", + "UMask": "0x80" + }, { "BriefDescription": "MEM_STORE_RETIRED.L2_HIT", "Counter": "0,1,2,3", "EventCode": "0x44", "EventName": "MEM_STORE_RETIRED.L2_HIT", - "PublicDescription": "MEM_STORE_RETIRED.L2_HIT Available PDIST cou= nters: 0", "SampleAfterValue": "200003", "UMask": "0x1" }, @@ -642,7 +657,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe5", "EventName": "MEM_UOP_RETIRED.ANY", - "PublicDescription": "Number of retired micro-operations (uops) fo= r load or store memory accesses Available PDIST counters: 0", + "PublicDescription": "Number of retired micro-operations (uops) fo= r load or store memory accesses", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -690,6 +705,17 @@ "SampleAfterValue": "100003", "UMask": "0x1" }, + { + "BriefDescription": "Counts demand data reads that were supplied b= y CXL MEM (Type 2 or Type 3).", + "Counter": "0,1,2,3", + "EventCode": "0x2A,0x2B", + "EventName": "OCR.DEMAND_DATA_RD.CXL_MEM", + "MSRIndex": "0x1a6,0x1a7", + "MSRValue": "0x703C00001", + "PublicDescription": "Counts demand data reads that were supplied = by CXL MEM (Type 2 or Type 3). Available PDIST counters: 0", + "SampleAfterValue": "100003", + "UMask": "0x1" + }, { "BriefDescription": "Counts demand data reads that hit in the L3 o= r were snooped from another core's caches on the same socket.", "Counter": "0,1,2,3", @@ -734,6 +760,17 @@ "SampleAfterValue": "100003", "UMask": "0x1" }, + { + "BriefDescription": "Counts demand data reads that were supplied b= y CXL MEM (Type 2 and Type 3) attached to local socket.", + "Counter": "0,1,2,3", + "EventCode": "0x2A,0x2B", + "EventName": "OCR.DEMAND_DATA_RD.LOCAL_CXL_MEM", + "MSRIndex": "0x1a6,0x1a7", + "MSRValue": "0x700C00001", + "PublicDescription": "Counts demand data reads that were supplied = by CXL MEM (Type 2 and Type 3) attached to local socket. Available PDIST co= unters: 0", + "SampleAfterValue": "100003", + "UMask": "0x1" + }, { "BriefDescription": "Counts demand data reads that were supplied b= y a cache on a remote socket where a snoop hit a modified line in another c= ore's caches which forwarded the data.", "Counter": "0,1,2,3", @@ -756,6 +793,17 @@ "SampleAfterValue": "100003", "UMask": "0x1" }, + { + "BriefDescription": "Counts demand data reads that were supplied b= y CXL MEM (Type 2 or Type 3) attached to another socket.", + "Counter": "0,1,2,3", + "EventCode": "0x2A,0x2B", + "EventName": "OCR.DEMAND_DATA_RD.REMOTE_CXL_MEM", + "MSRIndex": "0x1a6,0x1a7", + "MSRValue": "0x703000001", + "PublicDescription": "Counts demand data reads that were supplied = by CXL MEM (Type 2 or Type 3) attached to another socket. Available PDIST c= ounters: 0", + "SampleAfterValue": "100003", + "UMask": "0x1" + }, { "BriefDescription": "Counts demand data reads that hit a modified = line in a distant L3 Cache or were snooped from a distant core's L1/L2 cach= es on this socket when the system is in SNC (sub-NUMA cluster) mode.", "Counter": "0,1,2,3", @@ -789,6 +837,17 @@ "SampleAfterValue": "100003", "UMask": "0x1" }, + { + "BriefDescription": "Counts demand reads for ownership (RFO) reque= sts and software prefetches for exclusive ownership (PREFETCHW) that were s= upplied by CXL MEM (Type 2 or Type 3).", + "Counter": "0,1,2,3", + "EventCode": "0x2A,0x2B", + "EventName": "OCR.DEMAND_RFO.CXL_MEM", + "MSRIndex": "0x1a6,0x1a7", + "MSRValue": "0x703C00002", + "PublicDescription": "Counts demand reads for ownership (RFO) requ= ests and software prefetches for exclusive ownership (PREFETCHW) that were = supplied by CXL MEM (Type 2 or Type 3). Available PDIST counters: 0", + "SampleAfterValue": "100003", + "UMask": "0x1" + }, { "BriefDescription": "Counts demand reads for ownership (RFO) reque= sts and software prefetches for exclusive ownership (PREFETCHW) that hit in= the L3 or were snooped from another core's caches on the same socket.", "Counter": "0,1,2,3", @@ -811,6 +870,28 @@ "SampleAfterValue": "100003", "UMask": "0x1" }, + { + "BriefDescription": "Counts demand reads for ownership (RFO) reque= sts and software prefetches for exclusive ownership (PREFETCHW) that were s= upplied by CXL MEM (Type 2 and Type 3) attached to local socket.", + "Counter": "0,1,2,3", + "EventCode": "0x2A,0x2B", + "EventName": "OCR.DEMAND_RFO.LOCAL_CXL_MEM", + "MSRIndex": "0x1a6,0x1a7", + "MSRValue": "0x700C00002", + "PublicDescription": "Counts demand reads for ownership (RFO) requ= ests and software prefetches for exclusive ownership (PREFETCHW) that were = supplied by CXL MEM (Type 2 and Type 3) attached to local socket. Available= PDIST counters: 0", + "SampleAfterValue": "100003", + "UMask": "0x1" + }, + { + "BriefDescription": "Counts demand reads for ownership (RFO) reque= sts and software prefetches for exclusive ownership (PREFETCHW) that were s= upplied by CXL MEM (Type 2 or Type 3) attached to another socket.", + "Counter": "0,1,2,3", + "EventCode": "0x2A,0x2B", + "EventName": "OCR.DEMAND_RFO.REMOTE_CXL_MEM", + "MSRIndex": "0x1a6,0x1a7", + "MSRValue": "0x703000002", + "PublicDescription": "Counts demand reads for ownership (RFO) requ= ests and software prefetches for exclusive ownership (PREFETCHW) that were = supplied by CXL MEM (Type 2 or Type 3) attached to another socket. Availabl= e PDIST counters: 0", + "SampleAfterValue": "100003", + "UMask": "0x1" + }, { "BriefDescription": "Counts writebacks of modified cachelines and = streaming stores that have any type of response.", "Counter": "0,1,2,3", @@ -833,6 +914,17 @@ "SampleAfterValue": "100003", "UMask": "0x1" }, + { + "BriefDescription": "Counts all (cacheable) data read, code read a= nd RFO requests including demands and prefetches to the core caches (L1 or = L2) that were supplied by CXL MEM (Type 2 or Type 3).", + "Counter": "0,1,2,3", + "EventCode": "0x2A,0x2B", + "EventName": "OCR.READS_TO_CORE.CXL_MEM", + "MSRIndex": "0x1a6,0x1a7", + "MSRValue": "0x703C04477", + "PublicDescription": "Counts all (cacheable) data read, code read = and RFO requests including demands and prefetches to the core caches (L1 or= L2) that were supplied by CXL MEM (Type 2 or Type 3). Available PDIST coun= ters: 0", + "SampleAfterValue": "100003", + "UMask": "0x1" + }, { "BriefDescription": "Counts all (cacheable) data read, code read a= nd RFO requests including demands and prefetches to the core caches (L1 or = L2) that hit in the L3 or were snooped from another core's caches on the sa= me socket.", "Counter": "0,1,2,3", @@ -855,6 +947,17 @@ "SampleAfterValue": "100003", "UMask": "0x1" }, + { + "BriefDescription": "Counts all (cacheable) data read, code read a= nd RFO requests including demands and prefetches to the core caches (L1 or = L2) that were supplied by CXL MEM (Type 2 and Type 3) attached to local soc= ket.", + "Counter": "0,1,2,3", + "EventCode": "0x2A,0x2B", + "EventName": "OCR.READS_TO_CORE.LOCAL_CXL_MEM", + "MSRIndex": "0x1a6,0x1a7", + "MSRValue": "0x700C04477", + "PublicDescription": "Counts all (cacheable) data read, code read = and RFO requests including demands and prefetches to the core caches (L1 or= L2) that were supplied by CXL MEM (Type 2 and Type 3) attached to local so= cket. Available PDIST counters: 0", + "SampleAfterValue": "100003", + "UMask": "0x1" + }, { "BriefDescription": "Counts all (cacheable) data read, code read a= nd RFO requests including demands and prefetches to the core caches (L1 or = L2) that were not supplied by the local socket's L1, L2, or L3 caches and w= ere supplied by a remote socket.", "Counter": "0,1,2,3", @@ -899,6 +1002,17 @@ "SampleAfterValue": "100003", "UMask": "0x1" }, + { + "BriefDescription": "Counts all (cacheable) data read, code read a= nd RFO requests including demands and prefetches to the core caches (L1 or = L2) that were supplied by CXL MEM (Type 2 or Type 3) attached to another so= cket.", + "Counter": "0,1,2,3", + "EventCode": "0x2A,0x2B", + "EventName": "OCR.READS_TO_CORE.REMOTE_CXL_MEM", + "MSRIndex": "0x1a6,0x1a7", + "MSRValue": "0x703004477", + "PublicDescription": "Counts all (cacheable) data read, code read = and RFO requests including demands and prefetches to the core caches (L1 or= L2) that were supplied by CXL MEM (Type 2 or Type 3) attached to another s= ocket. Available PDIST counters: 0", + "SampleAfterValue": "100003", + "UMask": "0x1" + }, { "BriefDescription": "Counts all (cacheable) data read, code read a= nd RFO requests including demands and prefetches to the core caches (L1 or = L2) that hit a modified line in a distant L3 Cache or were snooped from a d= istant core's L1/L2 caches on this socket when the system is in SNC (sub-NU= MA cluster) mode.", "Counter": "0,1,2,3", @@ -937,7 +1051,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.ALL_REQUESTS", - "PublicDescription": "Counts memory transactions reached the super= queue including requests initiated by the core, all L3 prefetches, page wa= lks, etc.. Available PDIST counters: 0", + "PublicDescription": "Counts memory transactions reached the super= queue including requests initiated by the core, all L3 prefetches, page wa= lks, etc..", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -946,7 +1060,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DATA_RD", - "PublicDescription": "Counts the demand and prefetch data reads. A= ll Core Data Reads include cacheable 'Demands' and L2 prefetchers (not L3 p= refetchers). Counting also covers reads due to page walks resulted from any= request type. Available PDIST counters: 0", + "PublicDescription": "Counts the demand and prefetch data reads. A= ll Core Data Reads include cacheable 'Demands' and L2 prefetchers (not L3 p= refetchers). Counting also covers reads due to page walks resulted from any= request type.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -955,7 +1069,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_CODE_RD", - "PublicDescription": "Counts both cacheable and Non-Cacheable code= read requests. Available PDIST counters: 0", + "PublicDescription": "Counts both cacheable and Non-Cacheable code= read requests.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -964,7 +1078,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_DATA_RD", - "PublicDescription": "Counts the Demand Data Read requests sent to= uncore. Use it in conjunction with OFFCORE_REQUESTS_OUTSTANDING to determi= ne average latency in the uncore. Available PDIST counters: 0", + "PublicDescription": "Counts the Demand Data Read requests sent to= uncore. Use it in conjunction with OFFCORE_REQUESTS_OUTSTANDING to determi= ne average latency in the uncore.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -973,7 +1087,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_RFO", - "PublicDescription": "Counts the demand RFO (read for ownership) r= equests including regular RFOs, locks, ItoM. Available PDIST counters: 0", + "PublicDescription": "Counts the demand RFO (read for ownership) r= equests including regular RFOs, locks, ItoM.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -982,7 +1096,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.MEM_UC", - "PublicDescription": "This event counts noncacheable memory data r= ead transactions. Available PDIST counters: 0", + "PublicDescription": "This event counts noncacheable memory data r= ead transactions.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -992,7 +1106,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD", - "PublicDescription": "Counts cycles when offcore outstanding cache= able Core Data Read transactions are present in the super queue. A transact= ion is considered to be in the Offcore outstanding state between L2 miss an= d transaction completion sent to requestor (SQ de-allocation). See correspo= nding Umask under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when offcore outstanding cache= able Core Data Read transactions are present in the super queue. A transact= ion is considered to be in the Offcore outstanding state between L2 miss an= d transaction completion sent to requestor (SQ de-allocation). See correspo= nding Umask under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -1002,7 +1116,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_CODE= _RD", - "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1012,7 +1126,6 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA= _RD", - "PublicDescription": "Cycles where at least 1 outstanding demand d= ata read request is pending. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1022,7 +1135,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO", - "PublicDescription": "Counts the number of offcore outstanding dem= and rfo Reads transactions in the super queue every cycle. The 'Offcore out= standing' state of the transaction lasts from the L2 miss until the sending= transaction completion to requestor (SQ deallocation). See the correspondi= ng Umask under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding dem= and rfo Reads transactions in the super queue every cycle. The 'Offcore out= standing' state of the transaction lasts from the L2 miss until the sending= transaction completion to requestor (SQ deallocation). See the correspondi= ng Umask under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -1031,7 +1144,6 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DATA_RD", - "PublicDescription": "OFFCORE_REQUESTS_OUTSTANDING.DATA_RD Availab= le PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -1040,7 +1152,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_CODE_RD", - "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1049,7 +1161,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD", - "PublicDescription": "For every cycle, increments by the number of= outstanding demand data read requests pending. Requests are considered o= utstanding from the time they miss the core's L2 cache until the transactio= n completion message is sent to the requestor. Available PDIST counters: 0", + "PublicDescription": "For every cycle, increments by the number of= outstanding demand data read requests pending. Requests are considered o= utstanding from the time they miss the core's L2 cache until the transactio= n completion message is sent to the requestor.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -1058,7 +1170,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_RFO", - "PublicDescription": "Counts the number of off-core outstanding re= ad-for-ownership (RFO) store transactions every cycle. An RFO transaction i= s considered to be in the Off-core outstanding state between L2 cache miss = and transaction completion. Available PDIST counters: 0", + "PublicDescription": "Counts the number of off-core outstanding re= ad-for-ownership (RFO) store transactions every cycle. An RFO transaction i= s considered to be in the Off-core outstanding state between L2 cache miss = and transaction completion.", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -1067,7 +1179,7 @@ "Counter": "0,1,2,3", "EventCode": "0x2c", "EventName": "SQ_MISC.BUS_LOCK", - "PublicDescription": "Counts the more expensive bus lock needed to= enforce cache coherency for certain memory accesses that need to be done a= tomically. Can be created by issuing an atomic instruction (via the LOCK p= refix) which causes a cache line split or accesses uncacheable memory. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts the more expensive bus lock needed to= enforce cache coherency for certain memory accesses that need to be done a= tomically. Can be created by issuing an atomic instruction (via the LOCK p= refix) which causes a cache line split or accesses uncacheable memory.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -1076,7 +1188,6 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.ANY", - "PublicDescription": "Counts the number of PREFETCHNTA, PREFETCHW,= PREFETCHT0, PREFETCHT1 or PREFETCHT2 instructions executed. Available PDIS= T counters: 0", "SampleAfterValue": "100003", "UMask": "0xf" }, @@ -1085,7 +1196,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.NTA", - "PublicDescription": "Counts the number of PREFETCHNTA instruction= s executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHNTA instruction= s executed.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -1094,7 +1205,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.PREFETCHW", - "PublicDescription": "Counts the number of PREFETCHW instructions = executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHW instructions = executed.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -1103,7 +1214,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.T0", - "PublicDescription": "Counts the number of PREFETCHT0 instructions= executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHT0 instructions= executed.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -1112,7 +1223,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.T1_T2", - "PublicDescription": "Counts the number of PREFETCHT1 or PREFETCHT= 2 instructions executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHT1 or PREFETCHT= 2 instructions executed.", "SampleAfterValue": "100003", "UMask": "0x4" } diff --git a/tools/perf/pmu-events/arch/x86/graniterapids/floating-point.js= on b/tools/perf/pmu-events/arch/x86/graniterapids/floating-point.json index 1832dd952f66..59789eee060c 100644 --- a/tools/perf/pmu-events/arch/x86/graniterapids/floating-point.json +++ b/tools/perf/pmu-events/arch/x86/graniterapids/floating-point.json @@ -5,7 +5,6 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.FPDIV_ACTIVE", - "PublicDescription": "This event counts the cycles the floating po= int divider is busy. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -14,7 +13,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.FP", - "PublicDescription": "Counts all microcode Floating Point assists.= Available PDIST counters: 0", + "PublicDescription": "Counts all microcode Floating Point assists.= ", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -23,7 +22,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.SSE_AVX_MIX", - "PublicDescription": "ASSISTS.SSE_AVX_MIX Available PDIST counters= : 0", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -32,7 +30,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_0", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_0 [This event is al= ias to FP_ARITH_DISPATCHED.V0] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -41,7 +38,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_1", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_1 [This event is al= ias to FP_ARITH_DISPATCHED.V1] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -50,7 +46,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_5", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_5 [This event is al= ias to FP_ARITH_DISPATCHED.V2] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -59,7 +54,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V0", - "PublicDescription": "FP_ARITH_DISPATCHED.V0 [This event is alias = to FP_ARITH_DISPATCHED.PORT_0] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -68,7 +62,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V1", - "PublicDescription": "FP_ARITH_DISPATCHED.V1 [This event is alias = to FP_ARITH_DISPATCHED.PORT_1] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -77,7 +70,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V2", - "PublicDescription": "FP_ARITH_DISPATCHED.V2 [This event is alias = to FP_ARITH_DISPATCHED.PORT_5] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -86,7 +78,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 2 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as the= y perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR re= gister need to be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 2 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as the= y perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR re= gister need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -95,7 +87,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events. Available PDIST co= unters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -104,7 +96,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 = calculations per element. The DAZ and FTZ flags in the MXCSR register need = to be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 = calculations per element. The DAZ and FTZ flags in the MXCSR register need = to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -113,7 +105,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events. Available PDIST co= unters: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -122,7 +114,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.4_FLOPS", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision and 256-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 2 or/and 4 computation operations, one for each element. = Applies to SSE* and AVX* packed single precision floating-point and packed= double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL= DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB ins= tructions count twice as they perform 2 calculations per element. The DAZ a= nd FTZ flags in the MXCSR register need to be set when using these events. = Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision and 256-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 2 or/and 4 computation operations, one for each element. = Applies to SSE* and AVX* packed single precision floating-point and packed= double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL= DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB ins= tructions count twice as they perform 2 calculations per element. The DAZ a= nd FTZ flags in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x18" }, @@ -131,7 +123,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP14= FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calc= ulations per element. The DAZ and FTZ flags in the MXCSR register need to b= e set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP14= FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calc= ulations per element. The DAZ and FTZ flags in the MXCSR register need to b= e set when using these events.", "SampleAfterValue": "100003", "UMask": "0x40" }, @@ -140,7 +132,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 16 computation oper= ations, one for each element. Applies to SSE* and AVX* packed single preci= sion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP1= 4 FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 cal= culations per element. The DAZ and FTZ flags in the MXCSR register need to = be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 16 computation oper= ations, one for each element. Applies to SSE* and AVX* packed single preci= sion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP1= 4 FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 cal= culations per element. The DAZ and FTZ flags in the MXCSR register need to = be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -149,7 +141,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.8_FLOPS", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision and 512-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 8 computation operations, one for each element. Applies = to SSE* and AVX* packed single precision and double precision floating-poin= t instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RSQRT14= RCP RCP14 DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice= as they perform 2 calculations per element. The DAZ and FTZ flags in the M= XCSR register need to be set when using these events. Available PDIST count= ers: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision and 512-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 8 computation operations, one for each element. Applies = to SSE* and AVX* packed single precision and double precision floating-poin= t instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RSQRT14= RCP RCP14 DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice= as they perform 2 calculations per element. The DAZ and FTZ flags in the M= XCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x60" }, @@ -158,7 +150,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR", - "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision and double precision floating-point instructions retired; some = instructions will count twice as noted below. Each count represents 1 comp= utational operation. Applies to SSE* and AVX* scalar single precision float= ing-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB= . FM(N)ADD/SUB instructions count twice as they perform 2 calculations per= element. The DAZ and FTZ flags in the MXCSR register need to be set when u= sing these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision and double precision floating-point instructions retired; some = instructions will count twice as noted below. Each count represents 1 comp= utational operation. Applies to SSE* and AVX* scalar single precision float= ing-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB= . FM(N)ADD/SUB instructions count twice as they perform 2 calculations per= element. The DAZ and FTZ flags in the MXCSR register need to be set when u= sing these events.", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -167,7 +159,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational scalar doubl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar double precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions co= unt twice as they perform 2 calculations per element. The DAZ and FTZ flags= in the MXCSR register need to be set when using these events. Available PD= IST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar doubl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar double precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions co= unt twice as they perform 2 calculations per element. The DAZ and FTZ flags= in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -176,7 +168,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE", - "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar single precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instr= uctions count twice as they perform 2 calculations per element. The DAZ and= FTZ flags in the MXCSR register need to be set when using these events. Av= ailable PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar single precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instr= uctions count twice as they perform 2 calculations per element. The DAZ and= FTZ flags in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -185,7 +177,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.VECTOR", - "PublicDescription": "Number of any Vector retired FP arithmetic i= nstructions. The DAZ and FTZ flags in the MXCSR register need to be set wh= en using these events. Available PDIST counters: 0", + "PublicDescription": "Number of any Vector retired FP arithmetic i= nstructions. The DAZ and FTZ flags in the MXCSR register need to be set wh= en using these events.", "SampleAfterValue": "1000003", "UMask": "0xfc" }, @@ -194,7 +186,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.128B_PACKED_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.128B_PACKED_HALF Avai= lable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -203,7 +194,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.256B_PACKED_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.256B_PACKED_HALF Avai= lable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -212,7 +202,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.512B_PACKED_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.512B_PACKED_HALF Avai= lable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -221,7 +210,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.COMPLEX_SCALAR_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.COMPLEX_SCALAR_HALF A= vailable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -230,7 +218,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.SCALAR", - "PublicDescription": "FP_ARITH_INST_RETIRED2.SCALAR Available PDIS= T counters: 0", + "PublicDescription": "FP_ARITH_INST_RETIRED2.SCALAR", "SampleAfterValue": "100003", "UMask": "0x3" }, @@ -239,7 +227,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.SCALAR_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.SCALAR_HALF Available= PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -248,7 +235,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.VECTOR", - "PublicDescription": "FP_ARITH_INST_RETIRED2.VECTOR Available PDIS= T counters: 0", + "PublicDescription": "FP_ARITH_INST_RETIRED2.VECTOR", "SampleAfterValue": "100003", "UMask": "0x1c" } diff --git a/tools/perf/pmu-events/arch/x86/graniterapids/frontend.json b/t= ools/perf/pmu-events/arch/x86/graniterapids/frontend.json index b7cd92fbecd5..d580d305c926 100644 --- a/tools/perf/pmu-events/arch/x86/graniterapids/frontend.json +++ b/tools/perf/pmu-events/arch/x86/graniterapids/frontend.json @@ -4,7 +4,7 @@ "Counter": "0,1,2,3", "EventCode": "0x60", "EventName": "BACLEARS.ANY", - "PublicDescription": "Number of times the front-end is resteered w= hen it finds a branch instruction in a fetch line. This is called Unknown B= ranch which occurs for the first time a branch instruction is fetched or wh= en the branch is not tracked by the BPU (Branch Prediction Unit) anymore. A= vailable PDIST counters: 0", + "PublicDescription": "Number of times the front-end is resteered w= hen it finds a branch instruction in a fetch line. This is called Unknown B= ranch which occurs for the first time a branch instruction is fetched or wh= en the branch is not tracked by the BPU (Branch Prediction Unit) anymore.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -13,7 +13,7 @@ "Counter": "0,1,2,3", "EventCode": "0x87", "EventName": "DECODE.LCP", - "PublicDescription": "Counts cycles that the Instruction Length de= coder (ILD) stalls occurred due to dynamically changing prefix length of th= e decoded instruction (by operand size prefix instruction 0x66, address siz= e prefix instruction 0x67 or REX.W for Intel64). Count is proportional to t= he number of prefixes in a 16B-line. This may result in a three-cycle penal= ty for each LCP (Length changing prefix) in a 16-byte chunk. Available PDIS= T counters: 0", + "PublicDescription": "Counts cycles that the Instruction Length de= coder (ILD) stalls occurred due to dynamically changing prefix length of th= e decoded instruction (by operand size prefix instruction 0x66, address siz= e prefix instruction 0x67 or REX.W for Intel64). Count is proportional to t= he number of prefixes in a 16B-line. This may result in a three-cycle penal= ty for each LCP (Length changing prefix) in a 16-byte chunk.", "SampleAfterValue": "500009", "UMask": "0x1" }, @@ -22,7 +22,6 @@ "Counter": "0,1,2,3", "EventCode": "0x87", "EventName": "DECODE.MS_BUSY", - "PublicDescription": "Cycles the Microcode Sequencer is busy. Avai= lable PDIST counters: 0", "SampleAfterValue": "500009", "UMask": "0x2" }, @@ -31,7 +30,7 @@ "Counter": "0,1,2,3", "EventCode": "0x61", "EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES", - "PublicDescription": "Decode Stream Buffer (DSB) is a Uop-cache th= at holds translations of previously fetched instructions that were decoded = by the legacy x86 decode pipeline (MITE). This event counts fetch penalty c= ycles when a transition occurs from DSB to MITE. Available PDIST counters: = 0", + "PublicDescription": "Decode Stream Buffer (DSB) is a Uop-cache th= at holds translations of previously fetched instructions that were decoded = by the legacy x86 decode pipeline (MITE). This event counts fetch penalty c= ycles when a transition occurs from DSB to MITE.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -303,7 +302,7 @@ "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE_DATA.STALLS", - "PublicDescription": "Counts cycles where a code line fetch is sta= lled due to an L1 instruction cache miss. The decode pipeline works at a 32= Byte granularity. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where a code line fetch is sta= lled due to an L1 instruction cache miss. The decode pipeline works at a 32= Byte granularity.", "SampleAfterValue": "500009", "UMask": "0x4" }, @@ -314,7 +313,6 @@ "EdgeDetect": "1", "EventCode": "0x80", "EventName": "ICACHE_DATA.STALL_PERIODS", - "PublicDescription": "ICACHE_DATA.STALL_PERIODS Available PDIST co= unters: 0", "SampleAfterValue": "500009", "UMask": "0x4" }, @@ -323,7 +321,7 @@ "Counter": "0,1,2,3", "EventCode": "0x83", "EventName": "ICACHE_TAG.STALLS", - "PublicDescription": "Counts cycles where a code fetch is stalled = due to L1 instruction cache tag miss. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where a code fetch is stalled = due to L1 instruction cache tag miss.", "SampleAfterValue": "200003", "UMask": "0x4" }, @@ -333,7 +331,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES_ANY", - "PublicDescription": "Counts the number of cycles uops were delive= red to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) p= ath. Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles uops were delive= red to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) p= ath.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -343,7 +341,7 @@ "CounterMask": "6", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES_OK", - "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the D= SB (Decode Stream Buffer) path. Count includes uops that may 'bypass' the I= DQ. Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the D= SB (Decode Stream Buffer) path. Count includes uops that may 'bypass' the I= DQ.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -352,7 +350,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.DSB_UOPS", - "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Availab= le PDIST counters: 0", + "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -362,7 +360,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES_ANY", - "PublicDescription": "Counts the number of cycles uops were delive= red to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipe= line) path. During these cycles uops are not being delivered from the Decod= e Stream Buffer (DSB). Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles uops were delive= red to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipe= line) path. During these cycles uops are not being delivered from the Decod= e Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -372,7 +370,7 @@ "CounterMask": "6", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES_OK", - "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the M= ITE (legacy decode pipeline) path. During these cycles uops are not being d= elivered from the Decode Stream Buffer (DSB). Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the M= ITE (legacy decode pipeline) path. During these cycles uops are not being d= elivered from the Decode Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -381,7 +379,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_UOPS", - "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the MITE path. This also means that uops are= not being delivered from the Decode Stream Buffer (DSB). Available PDIST c= ounters: 0", + "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the MITE path. This also means that uops are= not being delivered from the Decode Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -391,7 +389,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_CYCLES_ANY", - "PublicDescription": "Counts cycles during which uops are being de= livered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS= ) is busy. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts cycles during which uops are being de= livered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS= ) is busy. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.", "SampleAfterValue": "2000003", "UMask": "0x20" }, @@ -402,7 +400,7 @@ "EdgeDetect": "1", "EventCode": "0x79", "EventName": "IDQ.MS_SWITCHES", - "PublicDescription": "Number of switches from DSB (Decode Stream B= uffer) or MITE (legacy decode pipeline) to the Microcode Sequencer. Availab= le PDIST counters: 0", + "PublicDescription": "Number of switches from DSB (Decode Stream B= uffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -411,7 +409,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_UOPS", - "PublicDescription": "Counts the number of uops initiated by MITE = or Decode Stream Buffer (DSB) and delivered to Instruction Decode Queue (ID= Q) while the Microcode Sequencer (MS) is busy. Counting includes uops that = may 'bypass' the IDQ. Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops initiated by MITE = or Decode Stream Buffer (DSB) and delivered to Instruction Decode Queue (ID= Q) while the Microcode Sequencer (MS) is busy. Counting includes uops that = may 'bypass' the IDQ.", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -420,7 +418,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CORE", - "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that when no operation was delivered to the back-end pipeline due= to instruction fetch limitations when the back-end could have accepted mor= e operations. Common examples include instruction cache misses or x86 instr= uction decode limitations. The count may be distributed among unhalted logi= cal processors (hyper-threads) who share the same physical core, in process= ors that support Intel Hyper-Threading Technology. Software can use this ev= ent as the numerator for the Frontend Bound metric (or top-level category) = of the Top-down Microarchitecture Analysis method. Available PDIST counters= : 0", + "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that when no operation was delivered to the back-end pipeline due= to instruction fetch limitations when the back-end could have accepted mor= e operations. Common examples include instruction cache misses or x86 instr= uction decode limitations. The count may be distributed among unhalted logi= cal processors (hyper-threads) who share the same physical core, in process= ors that support Intel Hyper-Threading Technology. Software can use this ev= ent as the numerator for the Frontend Bound metric (or top-level category) = of the Top-down Microarchitecture Analysis method.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -430,7 +428,7 @@ "CounterMask": "6", "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CYCLES_0_UOPS_DELIV.CORE", - "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CYCLES= _0_UOPS_DELIV.CORE] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CYCLES= _0_UOPS_DELIV.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -441,7 +439,7 @@ "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CYCLES_FE_WAS_OK", "Invert": "1", - "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_UOPS_N= OT_DELIVERED.CYCLES_FE_WAS_OK] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_UOPS_N= OT_DELIVERED.CYCLES_FE_WAS_OK]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -450,7 +448,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CORE", - "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -460,7 +458,7 @@ "CounterMask": "6", "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE", - "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_BUBBLES.CYCLES_0_UOPS_DEL= IV.CORE] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_BUBBLES.CYCLES_0_UOPS_DEL= IV.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -471,7 +469,7 @@ "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK", "Invert": "1", - "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_BUBBLE= S.CYCLES_FE_WAS_OK] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_BUBBLE= S.CYCLES_FE_WAS_OK]", "SampleAfterValue": "1000003", "UMask": "0x1" } diff --git a/tools/perf/pmu-events/arch/x86/graniterapids/gnr-metrics.json = b/tools/perf/pmu-events/arch/x86/graniterapids/gnr-metrics.json index 9a620e1b8de8..cc3c834ca286 100644 --- a/tools/perf/pmu-events/arch/x86/graniterapids/gnr-metrics.json +++ b/tools/perf/pmu-events/arch/x86/graniterapids/gnr-metrics.json @@ -1,28 +1,28 @@ [ { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" @@ -381,7 +381,7 @@ { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "topdown\\-be\\-bound / (topdown\\-fe\\-bound + topd= own\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_inf= o_thread_slots", + "MetricExpr": "topdown\\-be\\-bound / (topdown\\-fe\\-bound + topd= own\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BvOB;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", @@ -415,40 +415,40 @@ "MetricThreshold": "tma_bottleneck_branching_overhead > 5", "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_amx_busy= + tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_c= ore_bound * tma_amx_busy / (tma_amx_busy + tma_divider + tma_ports_utilizat= ion + tma_serializing_operation) + tma_core_bound * (tma_ports_utilization = / (tma_amx_busy + tma_divider + tma_ports_utilization + tma_serializing_ope= ration)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_utili= zed_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)= ))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_late= ncy)) + tma_memory_bound * (tma_l3_bound / (tma_cxl_mem_bound + tma_dram_bo= und + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma= _sq_full / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_latency = + tma_sq_full)) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + t= ma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_boun= d)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependen= cy + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependen= cy + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_= bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma= _l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_dtlb_load + tma_fb= _full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tm= a_store_fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tm= a_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_split_l= oads / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_= latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_s= tore_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_sharin= g + tma_split_stores + tma_store_latency + tma_streaming_stores)) + tma_mem= ory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_boun= d + tma_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store= + tma_false_sharing + tma_split_stores + tma_store_latency + tma_streaming= _stores)))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latenc= y)) + 0 / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound= + tma_l3_bound + tma_store_bound) * tma_mem_latency / (tma_mem_bandwidth += tma_mem_latency) + tma_memory_bound * (tma_l3_bound / (tma_cxl_mem_bound += tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bo= und)) * (tma_l3_hit_latency / (tma_contested_accesses + tma_data_sharing + = tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * tma_l2_bound / (tma= _cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_boun= d + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_boun= d + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store= _bound)) * (tma_l1_latency_dependency / (tma_dtlb_load + tma_fb_full + tma_= l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_= blk)) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + tma_dram_bo= und + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma= _lock_latency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + = tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound= * (tma_l1_bound / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma= _l2_bound + tma_l3_bound + tma_store_bound)) * (tma_split_loads / (tma_dtlb= _load + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_sp= lit_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tm= a_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_sh= aring + tma_split_stores + tma_store_latency + tma_streaming_stores)) + tma= _memory_bound * (tma_store_bound / (tma_cxl_mem_bound + tma_dram_bound + tm= a_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_store_l= atency / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_store= _latency + tma_streaming_stores)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_amx_busy= + tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_c= ore_bound * tma_amx_busy / (tma_amx_busy + tma_divider + tma_ports_utilizat= ion + tma_serializing_operation) + tma_core_bound * (tma_ports_utilization = / (tma_amx_busy + tma_divider + tma_ports_utilization + tma_serializing_ope= ration)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_utili= zed_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - I= NST_RETIRED.REP_ITERATION / cpu@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetc= h_latency * (tma_ms_switches + tma_branch_resteers * (tma_clears_resteers += tma_mispredicts_resteers * tma_other_mispredicts / tma_branch_mispredicts)= / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_branches))= / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_m= isses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * tma_ms / (tma_ds= b + tma_mite + tma_ms))) - tma_bottleneck_big_code", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - I= NST_RETIRED.REP_ITERATION / cpu@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetc= h_latency * (tma_ms_switches + tma_branch_resteers * (tma_clears_resteers += tma_mispredicts_resteers * tma_other_mispredicts / tma_branch_mispredicts)= / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_branches))= / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_m= isses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottleneck_big_code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20" }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * ((1 - INST_RETIRED.REP_ITERATION / cpu@UOPS_R= ETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_switches + tma_bra= nch_resteers * (tma_clears_resteers + tma_mispredicts_resteers * tma_other_= mispredicts / tma_branch_mispredicts) / (tma_clears_resteers + tma_mispredi= cts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_dsb_swit= ches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) + t= ma_fetch_bandwidth * tma_ms / (tma_dsb + tma_mite + tma_ms)) + 10 * tma_mic= rocode_sequencer * tma_other_mispredicts / tma_branch_mispredicts * tma_bra= nch_mispredicts + tma_machine_clears * tma_other_nukes / tma_other_nukes + = tma_core_bound * (tma_serializing_operation + RS.EMPTY_RESOURCE / tma_info_= thread_clks * tma_ports_utilized_0) / (tma_amx_busy + tma_divider + tma_por= ts_utilization + tma_serializing_operation) + tma_microcode_sequencer / (tm= a_few_uops_instructions + tma_microcode_sequencer) * (tma_assists / tma_mic= rocode_sequencer) * tma_heavy_operations)", + "MetricExpr": "100 * ((1 - INST_RETIRED.REP_ITERATION / cpu@UOPS_R= ETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_switches + tma_bra= nch_resteers * (tma_clears_resteers + tma_mispredicts_resteers * tma_other_= mispredicts / tma_branch_mispredicts) / (tma_clears_resteers + tma_mispredi= cts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_dsb_swit= ches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) + t= ma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_= mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_nukes= / tma_other_nukes + tma_core_bound * (tma_serializing_operation + RS.EMPTY= _RESOURCE / tma_info_thread_clks * tma_ports_utilized_0) / (tma_amx_busy + = tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_micr= ocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) * (= tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -456,7 +456,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", - "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / (tma_dram= _bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (= tma_dtlb_load / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + = tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound= * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l= 3_bound + tma_store_bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_false= _sharing + tma_split_stores + tma_store_latency + tma_streaming_stores)))", + "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / (tma_cxl_= mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + t= ma_store_bound)) * (tma_dtlb_load / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)= ) + tma_memory_bound * (tma_store_bound / (tma_cxl_mem_bound + tma_dram_bou= nd + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_= dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_s= tore_latency + tma_streaming_stores)))", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", "MetricThreshold": "tma_bottleneck_memory_data_tlbs > 20", @@ -464,7 +464,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Synchronization= related bottlenecks (data transfers and coherency updates across processor= s)", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * = (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) * tma_remote_cach= e / (tma_local_mem + tma_remote_cache + tma_remote_mem) + tma_l3_bound / (t= ma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_boun= d) * (tma_contested_accesses + tma_data_sharing) / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full) + tma_store_bound / = (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bo= und) * tma_false_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_= stores + tma_store_latency + tma_streaming_stores - tma_store_latency)) + t= ma_machine_clears * (1 - tma_other_nukes / tma_other_nukes))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound) * (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency= )) * tma_remote_cache / (tma_local_mem + tma_remote_cache + tma_remote_mem)= + tma_l3_bound / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_= l2_bound + tma_l3_bound + tma_store_bound) * (tma_contested_accesses + tma_= data_sharing) / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_lat= ency + tma_sq_full) + tma_store_bound / (tma_cxl_mem_bound + tma_dram_bound= + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * tma_fals= e_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_st= ore_latency + tma_streaming_stores - tma_store_latency)) + tma_machine_clea= rs * (1 - tma_other_nukes / tma_other_nukes))", "MetricGroup": "BvMS;LockCont;Mem;Offcore;tma_issueSyncxn", "MetricName": "tma_bottleneck_memory_synchronization", "MetricThreshold": "tma_bottleneck_memory_synchronization > 10", @@ -480,7 +480,7 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -496,7 +496,7 @@ { "BriefDescription": "This metric represents fraction of slots the = CPU has wasted due to Branch Misprediction", "DefaultMetricgroupName": "TopdownL2", - "MetricExpr": "topdown\\-br\\-mispredict / (topdown\\-fe\\-bound += topdown\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tm= a_info_thread_slots", + "MetricExpr": "topdown\\-br\\-mispredict / (topdown\\-fe\\-bound += topdown\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BadSpec;BrMispredicts;BvMP;Default;TmaL2;TopdownL2= ;tma_L2_group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_specula= tion > 0.15", @@ -613,7 +613,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to contested acces= ses", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS * min(MEM_LOAD_L= 3_HIT_RETIRED.XSNP_MISS:R, 74.6 * tma_info_system_core_frequency) + MEM_LOA= D_L3_HIT_RETIRED.XSNP_FWD * min(MEM_LOAD_L3_HIT_RETIRED.XSNP_FWD:R, 76.6 * = tma_info_system_core_frequency) * (OCR.DEMAND_DATA_RD.L3_HIT.SNOOP_HITM / (= OCR.DEMAND_DATA_RD.L3_HIT.SNOOP_HITM + OCR.DEMAND_DATA_RD.L3_HIT.SNOOP_HIT_= WITH_FWD))) * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS / 2) = / tma_info_thread_clks", "MetricGroup": "BvMS;DataSharing;LockCont;Offcore;Snoop;TopdownL4;= tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", @@ -632,6 +631,15 @@ "PublicDescription": "This metric represents fraction of slots whe= re Core non-memory issues were of a bottleneck. Shortage in hardware compu= te resources; or dependencies in software's instructions are both categoriz= ed under Core Bound. Hence it may indicate the machine ran out of an out-of= -order resource; certain execution units are overloaded or dependencies in = program's data- or instruction-flow are limiting the performance (e.g. FP-c= hained long-latency arithmetic operations).", "ScaleUnit": "100%" }, + { + "BriefDescription": "This metric roughly estimates (based on idle = latencies) how often the CPU was stalled on accesses to external CXL Memory= by loads (e.g", + "MetricExpr": "(((1 - ((19 * (MEM_LOAD_L3_MISS_RETIRED.REMOTE_DRAM= * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS)) + 10 * (MEM_LO= AD_L3_MISS_RETIRED.LOCAL_DRAM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RET= IRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REMOTE_FWD * (1 + MEM_LOAD_RETIRED= .FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REMOTE_HITM = * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS))) / (19 * (MEM_L= OAD_L3_MISS_RETIRED.REMOTE_DRAM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_R= ETIRED.L1_MISS)) + 10 * (MEM_LOAD_L3_MISS_RETIRED.LOCAL_DRAM * (1 + MEM_LOA= D_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REM= OTE_FWD * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LO= AD_L3_MISS_RETIRED.REMOTE_HITM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RE= TIRED.L1_MISS)) + (25 * (MEM_LOAD_RETIRED.LOCAL_CXL_MEM * (1 + MEM_LOAD_RET= IRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS) if #has_pmem > 0 else 0) + 33 * (ME= M_LOAD_L3_MISS_RETIRED.REMOTE_CXL_MEM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_= LOAD_RETIRED.L1_MISS) if #has_pmem > 0 else 0))) if #has_pmem > 0 else 1)) = * (MEMORY_ACTIVITY.STALLS_L3_MISS / tma_info_thread_clks) if 1e6 * (MEM_LOA= D_L3_MISS_RETIRED.REMOTE_CXL_MEM + MEM_LOAD_RETIRED.LOCAL_CXL_MEM) > MEM_LO= AD_RETIRED.L1_MISS else 0) if #has_pmem > 0 else 0)", + "MetricGroup": "MemoryBound;Server;TmaL3mem;TopdownL3;tma_L3_group= ;tma_memory_bound_group", + "MetricName": "tma_cxl_mem_bound", + "MetricThreshold": "tma_cxl_mem_bound > 0.1 & (tma_memory_bound > = 0.2 & tma_backend_bound > 0.2)", + "PublicDescription": "This metric roughly estimates (based on idle= latencies) how often the CPU was stalled on accesses to external CXL Memor= y by loads (e.g. 3D-Xpoint (Crystal Ridge, a.k.a. IXP) memory, PMM - Persis= tent Memory Module [from CLX to SPR] or any other CXL Type3 Memory [EMR onw= ards]).", + "ScaleUnit": "100%" + }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to data-sharing ac= cesses", "MetricConstraint": "NO_GROUP_EVENTS", @@ -662,7 +670,7 @@ }, { "BriefDescription": "This metric estimates how often the CPU was s= talled on accesses to external memory (DRAM) by loads", - "MetricExpr": "MEMORY_ACTIVITY.STALLS_L3_MISS / tma_info_thread_cl= ks", + "MetricExpr": "(MEMORY_ACTIVITY.STALLS_L3_MISS / tma_info_thread_c= lks - tma_cxl_mem_bound if #has_pmem > 0 else MEMORY_ACTIVITY.STALLS_L3_MIS= S / tma_info_thread_clks)", "MetricGroup": "MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_me= mory_bound_group", "MetricName": "tma_dram_bound", "MetricThreshold": "tma_dram_bound > 0.1 & (tma_memory_bound > 0.2= & tma_backend_bound > 0.2)", @@ -720,7 +728,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%" }, { @@ -848,7 +856,7 @@ { "BriefDescription": "This metric represents fraction of slots wher= e the CPU was retiring heavy-weight operations -- instructions that require= two or more uops or micro-coded sequences", "DefaultMetricgroupName": "TopdownL2", - "MetricExpr": "topdown\\-heavy\\-ops / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_in= fo_thread_slots", + "MetricExpr": "topdown\\-heavy\\-ops / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "Default;Retire;TmaL2;TopdownL2;tma_L2_group;tma_re= tiring_group", "MetricName": "tma_heavy_operations", "MetricThreshold": "tma_heavy_operations > 0.1", @@ -1395,19 +1403,19 @@ { "BriefDescription": "Off-core accesses per kilo instruction for mo= dified write requests", "MetricExpr": "1e3 * OCR.MODIFIED_WRITE.ANY_RESPONSE / tma_info_in= st_mix_instructions", - "MetricGroup": "Offcore", + "MetricGroup": "Offcore;Server", "MetricName": "tma_info_memory_mix_offcore_mwrite_any_pki" }, { "BriefDescription": "Off-core accesses per kilo instruction for re= ads-to-core requests (speculative; including in-core HW prefetches)", "MetricExpr": "1e3 * OCR.READS_TO_CORE.ANY_RESPONSE / tma_info_ins= t_mix_instructions", - "MetricGroup": "CacheHits;Offcore", + "MetricGroup": "CacheHits;Offcore;Server", "MetricName": "tma_info_memory_mix_offcore_read_any_pki" }, { "BriefDescription": "L3 cache misses per kilo instruction for read= s-to-core requests (speculative; including in-core HW prefetches)", "MetricExpr": "1e3 * OCR.READS_TO_CORE.L3_MISS / tma_info_inst_mix= _instructions", - "MetricGroup": "Offcore", + "MetricGroup": "Offcore;Server", "MetricName": "tma_info_memory_mix_offcore_read_l3m_pki" }, { @@ -1433,21 +1441,21 @@ { "BriefDescription": "Average DRAM BW for Reads-to-Core (R2C) cover= ing for memory attached to local- and remote-socket", "MetricExpr": "64 * OCR.READS_TO_CORE.DRAM / 1e9 / tma_info_system= _time", - "MetricGroup": "HPC;Mem;MemoryBW;SoC", + "MetricGroup": "HPC;Mem;MemoryBW;Offcore;Server", "MetricName": "tma_info_memory_soc_r2c_dram_bw", "PublicDescription": "Average DRAM BW for Reads-to-Core (R2C) cove= ring for memory attached to local- and remote-socket. See R2C_Offcore_BW." }, { "BriefDescription": "Average L3-cache miss BW for Reads-to-Core (R= 2C)", "MetricExpr": "64 * OCR.READS_TO_CORE.L3_MISS / 1e9 / tma_info_sys= tem_time", - "MetricGroup": "HPC;Mem;MemoryBW;SoC", + "MetricGroup": "HPC;Mem;MemoryBW;Offcore;Server", "MetricName": "tma_info_memory_soc_r2c_l3m_bw", "PublicDescription": "Average L3-cache miss BW for Reads-to-Core (= R2C). This covering going to DRAM or other memory off-chip memory tears. Se= e R2C_Offcore_BW." }, { "BriefDescription": "Average Off-core access BW for Reads-to-Core = (R2C)", "MetricExpr": "64 * OCR.READS_TO_CORE.ANY_RESPONSE / 1e9 / tma_inf= o_system_time", - "MetricGroup": "HPC;Mem;MemoryBW;SoC", + "MetricGroup": "HPC;Mem;MemoryBW;Offcore;Server", "MetricName": "tma_info_memory_soc_r2c_offcore_bw", "PublicDescription": "Average Off-core access BW for Reads-to-Core= (R2C). R2C account for demand or prefetch load/RFO/code access that fill d= ata into the Core caches." }, @@ -1491,7 +1499,7 @@ "MetricName": "tma_info_memory_tlb_store_stlb_mpki" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (UOPS_EXECUTED.CORE_CYCLES_G= E_1 / 2 if #SMT_on else cpu@UOPS_EXECUTED.THREAD\\,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -1538,7 +1546,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -1550,16 +1558,28 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, + { + "BriefDescription": "Average 3DXP Memory Bandwidth Use for reads [= GB / sec]", + "MetricExpr": "(64 * UNC_CXLCM_RxC_PACK_BUF_INSERTS.MEM_DATA / 1e9= / tma_info_system_time if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryBW;Server;SoC", + "MetricName": "tma_info_system_cxl_mem_read_bw" + }, + { + "BriefDescription": "Average 3DXP Memory Bandwidth Use for Writes = [GB / sec]", + "MetricExpr": "(64 * UNC_CXLDP_TxC_AGF_INSERTS.M2S_DATA / 1e9 / tm= a_info_system_time if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryBW;Server;SoC", + "MetricName": "tma_info_system_cxl_mem_write_bw" + }, { "BriefDescription": "Average external Memory Bandwidth Use for rea= ds and writes [GB / sec]", "MetricExpr": "64 * (UNC_M_CAS_COUNT_SCH0.RD + UNC_M_CAS_COUNT_SCH= 1.RD + UNC_M_CAS_COUNT_SCH0.WR + UNC_M_CAS_COUNT_SCH1.WR) / 1e9 / tma_info_= system_time", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Giga Floating Point Operations Per Second", @@ -1771,12 +1791,12 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", "MetricExpr": "min(2 * (MEM_INST_RETIRED.ALL_LOADS - MEM_LOAD_RETI= RED.FB_HIT - MEM_LOAD_RETIRED.L1_MISS) * 20 / 100, max(CYCLE_ACTIVITY.CYCLE= S_MEM_ANY - MEMORY_ACTIVITY.CYCLES_L1D_MISS, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%" }, { @@ -1790,7 +1810,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles wit= h demand load accesses that hit the L2 cache under unloaded scenarios (poss= ibly L2 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "MEM_LOAD_RETIRED.L2_HIT * min(MEM_LOAD_RETIRED.L2_H= IT:R, 4.4 * tma_info_system_core_frequency) * (1 + MEM_LOAD_RETIRED.FB_HIT = / MEM_LOAD_RETIRED.L1_MISS / 2) / tma_info_thread_clks", "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_l2_bound_grou= p", "MetricName": "tma_l2_hit_latency", @@ -1809,12 +1828,11 @@ }, { "BriefDescription": "This metric estimates fraction of cycles with= demand load accesses that hit the L3 cache under unloaded scenarios (possi= bly L3 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "MEM_LOAD_RETIRED.L3_HIT * min(MEM_LOAD_RETIRED.L3_H= IT:R, 32.6 * tma_info_system_core_frequency) * (1 + MEM_LOAD_RETIRED.FB_HIT= / MEM_LOAD_RETIRED.L1_MISS / 2) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%" }, { @@ -1897,6 +1915,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU spent handling cache misses due to lock operations", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "MEM_INST_RETIRED.LOCK_LOADS * MEM_INST_RETIRED.LOCK= _LOADS:R / tma_info_thread_clks", "MetricGroup": "LockCont;Offcore;TopdownL4;tma_L4_group;tma_issueR= FO;tma_l1_bound_group", "MetricName": "tma_lock_latency", @@ -1929,7 +1948,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%" }, { @@ -1938,13 +1957,13 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%" }, { "BriefDescription": "This metric represents fraction of slots the = Memory subsystem within the Backend was a bottleneck", "DefaultMetricgroupName": "TopdownL2", - "MetricExpr": "topdown\\-mem\\-bound / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_in= fo_thread_slots", + "MetricExpr": "topdown\\-mem\\-bound / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "Backend;Default;TmaL2;TopdownL2;tma_L2_group;tma_b= ackend_bound_group", "MetricName": "tma_memory_bound", "MetricThreshold": "tma_memory_bound > 0.2 & tma_backend_bound > 0= .2", @@ -1954,7 +1973,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to LFENCE Instructions.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * MISC2_RETIRED.LFENCE / tma_info_thread_clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_memory_fence", @@ -2007,7 +2025,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "max(IDQ.MS_CYCLES_ANY, cpu@UOPS_RETIRED.MS\\,cmask\= \=3D1@ / (UOPS_RETIRED.SLOTS / UOPS_ISSUED.ANY)) / tma_info_core_core_clks = / 2", + "MetricExpr": "max(IDQ.MS_CYCLES_ANY, cpu@UOPS_RETIRED.MS\\,cmask\= \=3D1@ / (UOPS_RETIRED.SLOTS / UOPS_ISSUED.ANY)) / tma_info_core_core_clks = / 2.4", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -2042,6 +2060,7 @@ }, { "BriefDescription": "This metric represents the remaining light uo= ps fraction the CPU has executed - remaining means not covered by other sib= ling nodes", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_light_operations - (tma_fp_arith + tma_i= nt_operations + tma_memory_operations + tma_fused_instructions + tma_non_fu= sed_branches))", "MetricGroup": "Pipeline;TopdownL3;tma_L3_group;tma_light_operatio= ns_group", "MetricName": "tma_other_light_ops", @@ -2103,6 +2122,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "((tma_ports_utilized_0 * tma_info_thread_clks + (EX= E_ACTIVITY.1_PORTS_UTIL + tma_retiring * EXE_ACTIVITY.2_3_PORTS_UTIL)) / tm= a_info_thread_clks if ARITH.DIV_ACTIVE < CYCLE_ACTIVITY.STALLS_TOTAL - EXE_= ACTIVITY.BOUND_ON_LOADS else (EXE_ACTIVITY.1_PORTS_UTIL + tma_retiring * EX= E_ACTIVITY.2_3_PORTS_UTIL) / tma_info_thread_clks)", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", @@ -2112,6 +2132,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "max(EXE_ACTIVITY.EXE_BOUND_0_PORTS - RESOURCE_STALL= S.SCOREBOARD, 0) / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", @@ -2121,6 +2142,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "EXE_ACTIVITY.1_PORTS_UTIL / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", @@ -2130,7 +2152,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "EXE_ACTIVITY.2_PORTS_UTIL / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", @@ -2140,7 +2161,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 3 or more uops per cycle on all execution ports (Logical= Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "UOPS_EXECUTED.CYCLES_GE_3 / tma_info_thread_clks", "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_ut= ilization_group", "MetricName": "tma_ports_utilized_3m", @@ -2150,7 +2170,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling loads from remote cache in other socket= s including synchronizations issues", - "MetricExpr": "(MEM_LOAD_L3_MISS_RETIRED.REMOTE_HITM * MEM_LOAD_L3= _MISS_RETIRED.REMOTE_HITM:R + MEM_LOAD_L3_MISS_RETIRED.REMOTE_FWD * MEM_LOA= D_L3_MISS_RETIRED.REMOTE_FWD:R) * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_R= ETIRED.L1_MISS / 2) / tma_info_thread_clks", + "MetricExpr": "(MEM_LOAD_L3_MISS_RETIRED.REMOTE_HITM * PEBS + MEM_= LOAD_L3_MISS_RETIRED.REMOTE_FWD * PEBS) * (1 + MEM_LOAD_RETIRED.FB_HIT / ME= M_LOAD_RETIRED.L1_MISS / 2) / tma_info_thread_clks", "MetricGroup": "Offcore;Server;Snoop;TopdownL5;tma_L5_group;tma_is= sueSyncxn;tma_mem_latency_group", "MetricName": "tma_remote_cache", "MetricThreshold": "tma_remote_cache > 0.05 & (tma_mem_latency > 0= .1 & (tma_dram_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > = 0.2)))", @@ -2159,7 +2179,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling loads from remote memory", - "MetricExpr": "MEM_LOAD_L3_MISS_RETIRED.REMOTE_DRAM * MEM_LOAD_L3_= MISS_RETIRED.REMOTE_DRAM:R * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRE= D.L1_MISS / 2) / tma_info_thread_clks", + "MetricExpr": "MEM_LOAD_L3_MISS_RETIRED.REMOTE_DRAM * PEBS * (1 + = MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS / 2) / tma_info_thread_c= lks", "MetricGroup": "Server;Snoop;TopdownL5;tma_L5_group;tma_mem_latenc= y_group", "MetricName": "tma_remote_mem", "MetricThreshold": "tma_remote_mem > 0.1 & (tma_mem_latency > 0.1 = & (tma_dram_bound > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2= )))", @@ -2177,7 +2197,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_info_= thread_slots", + "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -2205,7 +2225,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to PAUSE Instructions", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "CPU_CLK_UNHALTED.PAUSE / tma_info_thread_clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_slow_pause", @@ -2237,7 +2256,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%" }, { diff --git a/tools/perf/pmu-events/arch/x86/graniterapids/memory.json b/too= ls/perf/pmu-events/arch/x86/graniterapids/memory.json index 4db39f304c2c..96f40390becf 100644 --- a/tools/perf/pmu-events/arch/x86/graniterapids/memory.json +++ b/tools/perf/pmu-events/arch/x86/graniterapids/memory.json @@ -5,7 +5,6 @@ "CounterMask": "2", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L3_MISS", - "PublicDescription": "Cycles while L3 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -15,7 +14,6 @@ "CounterMask": "6", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L3_MISS", - "PublicDescription": "Execution stalls while L3 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x6" }, @@ -24,7 +22,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.MEMORY_ORDERING", - "PublicDescription": "Counts the number of Machine Clears detected= dye to memory ordering. Memory Ordering Machine Clears may apply when a me= mory read may not conform to the memory ordering rules of the x86 architect= ure Available PDIST counters: 0", + "PublicDescription": "Counts the number of Machine Clears detected= dye to memory ordering. Memory Ordering Machine Clears may apply when a me= mory read may not conform to the memory ordering rules of the x86 architect= ure", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -34,7 +32,6 @@ "CounterMask": "2", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.CYCLES_L1D_MISS", - "PublicDescription": "Cycles while L1 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -44,7 +41,6 @@ "CounterMask": "3", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L1D_MISS", - "PublicDescription": "Execution stalls while L1 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -54,7 +50,7 @@ "CounterMask": "5", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L2_MISS", - "PublicDescription": "Execution stalls while L2 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock). Available PDIST counters: 0", + "PublicDescription": "Execution stalls while L2 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock).", "SampleAfterValue": "1000003", "UMask": "0x5" }, @@ -64,7 +60,7 @@ "CounterMask": "9", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L3_MISS", - "PublicDescription": "Execution stalls while L3 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock). Available PDIST counters: 0", + "PublicDescription": "Execution stalls while L3 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock).", "SampleAfterValue": "1000003", "UMask": "0x9" }, @@ -412,7 +408,6 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.L3_MISS_DEMAND_DATA_RD", - "PublicDescription": "Counts demand data read requests that miss t= he L3 cache. Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -422,7 +417,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_L3_MISS_DEM= AND_DATA_RD", - "PublicDescription": "Cycles with at least 1 Demand Data Read requ= ests who miss L3 cache in the superQ. Available PDIST counters: 0", + "PublicDescription": "Cycles with at least 1 Demand Data Read requ= ests who miss L3 cache in the superQ.", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -431,7 +426,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.L3_MISS_DEMAND_DATA_RD", - "PublicDescription": "For every cycle, increments by the number of= demand data read requests pending that are known to have missed the L3 cac= he. Note that this does not capture all elapsed cycles while requests are = outstanding - only cycles from when the requests were known by the requesti= ng core to have missed the L3 cache. Available PDIST counters: 0", + "PublicDescription": "For every cycle, increments by the number of= demand data read requests pending that are known to have missed the L3 cac= he. Note that this does not capture all elapsed cycles while requests are = outstanding - only cycles from when the requests were known by the requesti= ng core to have missed the L3 cache.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -449,7 +444,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_EVENTS", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to none of the previous 3 categories (e.g. interrupt). Availabl= e PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to none of the previous 3 categories (e.g. interrupt).", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -458,7 +453,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MEM", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to various memory events (e.g. read/write capacity and conflict= s). Available PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to various memory events (e.g. read/write capacity and conflict= s).", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -467,7 +462,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MEMTYPE", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to incompatible memory type. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to incompatible memory type.", "SampleAfterValue": "100003", "UMask": "0x40" }, @@ -476,7 +471,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_UNFRIENDLY", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to HLE-unfriendly instructions. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to HLE-unfriendly instructions.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -485,7 +480,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.COMMIT", - "PublicDescription": "Counts the number of times RTM commit succee= ded. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times RTM commit succee= ded.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -494,7 +489,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.START", - "PublicDescription": "Counts the number of times we entered an RTM= region. Does not count nested transactions. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times we entered an RTM= region. Does not count nested transactions.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -503,7 +498,7 @@ "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CAPACITY_READ", - "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional reads Available PDIST counters: 0", + "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional reads", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -512,7 +507,7 @@ "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CAPACITY_WRITE", - "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional writes. Available PDIST counters: 0", + "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional writes.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -521,7 +516,7 @@ "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CONFLICT", - "PublicDescription": "Counts the number of times a TSX line had a = cache conflict. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times a TSX line had a = cache conflict.", "SampleAfterValue": "100003", "UMask": "0x1" } diff --git a/tools/perf/pmu-events/arch/x86/graniterapids/other.json b/tool= s/perf/pmu-events/arch/x86/graniterapids/other.json index 8b7aa4caec46..c0747750b1a8 100644 --- a/tools/perf/pmu-events/arch/x86/graniterapids/other.json +++ b/tools/perf/pmu-events/arch/x86/graniterapids/other.json @@ -4,7 +4,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.HARDWARE", - "PublicDescription": "Count all other hardware assists or traps th= at are not necessarily architecturally exposed (through a software handler)= beyond FP; SSE-AVX mix and A/D assists who are counted by dedicated sub-ev= ents. This includes, but not limited to, assists at EXE or MEM uop writeba= ck like AVX* load/store/gather/scatter (non-FP GSSE-assist ) , assists gene= rated by ROB like PEBS and RTIT, Uncore trap, RAR (Remote Action Request) a= nd CET (Control flow Enforcement Technology) assists. the event also counts= for Machine Ordering count. Available PDIST counters: 0", + "PublicDescription": "Count all other hardware assists or traps th= at are not necessarily architecturally exposed (through a software handler)= beyond FP; SSE-AVX mix and A/D assists who are counted by dedicated sub-ev= ents. This includes, but not limited to, assists at EXE or MEM uop writeba= ck like AVX* load/store/gather/scatter (non-FP GSSE-assist ) , assists gene= rated by ROB like PEBS and RTIT, Uncore trap, RAR (Remote Action Request) a= nd CET (Control flow Enforcement Technology) assists. the event also counts= for Machine Ordering count.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -13,10 +13,34 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.PAGE_FAULT", - "PublicDescription": "ASSISTS.PAGE_FAULT Available PDIST counters:= 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, + { + "BriefDescription": "HW_INTERRUPTS.MASKED", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xcb", + "EventName": "HW_INTERRUPTS.MASKED", + "SampleAfterValue": "100003", + "UMask": "0x2" + }, + { + "BriefDescription": "HW_INTERRUPTS.PENDING_AND_MASKED", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xcb", + "EventName": "HW_INTERRUPTS.PENDING_AND_MASKED", + "SampleAfterValue": "100003", + "UMask": "0x4" + }, + { + "BriefDescription": "Number of hardware interrupts received by the= processor.", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xcb", + "EventName": "HW_INTERRUPTS.RECEIVED", + "PublicDescription": "Counts the number of hardware interruptions = received by the processor.", + "SampleAfterValue": "203", + "UMask": "0x1" + }, { "BriefDescription": "Counts streaming stores that have any type of= response.", "Counter": "0,1,2,3", @@ -34,7 +58,7 @@ "CounterMask": "1", "EventCode": "0x2d", "EventName": "XQ.FULL_CYCLES", - "PublicDescription": "number of cycles when the thread is active a= nd the uncore cannot take any further requests (for example prefetches, loa= ds or stores initiated by the Core that miss the L2 cache). Available PDIST= counters: 0", + "PublicDescription": "number of cycles when the thread is active a= nd the uncore cannot take any further requests (for example prefetches, loa= ds or stores initiated by the Core that miss the L2 cache).", "SampleAfterValue": "1000003", "UMask": "0x1" } diff --git a/tools/perf/pmu-events/arch/x86/graniterapids/pipeline.json b/t= ools/perf/pmu-events/arch/x86/graniterapids/pipeline.json index 27af3bd6bacf..0fef8fd61974 100644 --- a/tools/perf/pmu-events/arch/x86/graniterapids/pipeline.json +++ b/tools/perf/pmu-events/arch/x86/graniterapids/pipeline.json @@ -5,7 +5,7 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.DIV_ACTIVE", - "PublicDescription": "Counts cycles when divide unit is busy execu= ting divide or square root operations. Accounts for integer and floating-po= int operations. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when divide unit is busy execu= ting divide or square root operations. Accounts for integer and floating-po= int operations.", "SampleAfterValue": "1000003", "UMask": "0x9" }, @@ -15,7 +15,6 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.IDIV_ACTIVE", - "PublicDescription": "This event counts the cycles the integer div= ider is busy. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -24,7 +23,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.ANY", - "PublicDescription": "Counts the number of occurrences where a mic= rocode assist is invoked by hardware. Examples include AD (page Access Dirt= y), FP and AVX related assists. Available PDIST counters: 0", + "PublicDescription": "Counts the number of occurrences where a mic= rocode assist is invoked by hardware. Examples include AD (page Access Dirt= y), FP and AVX related assists.", "SampleAfterValue": "100003", "UMask": "0x1b" }, @@ -271,7 +270,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C01", - "PublicDescription": "Counts core clocks when the thread is in the= C0.1 light-weight slower wakeup time but more power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions. Availab= le PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.1 light-weight slower wakeup time but more power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -280,7 +279,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C02", - "PublicDescription": "Counts core clocks when the thread is in the= C0.2 light-weight faster wakeup time but less power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions. Availab= le PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.2 light-weight faster wakeup time but less power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions.", "SampleAfterValue": "2000003", "UMask": "0x20" }, @@ -289,7 +288,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C0_WAIT", - "PublicDescription": "Counts core clocks when the thread is in the= C0.1 or C0.2 power saving optimized states (TPAUSE or UMWAIT instructions)= or running the PAUSE instruction. Available PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.1 or C0.2 power saving optimized states (TPAUSE or UMWAIT instructions)= or running the PAUSE instruction.", "SampleAfterValue": "2000003", "UMask": "0x70" }, @@ -298,7 +297,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.DISTRIBUTED", - "PublicDescription": "This event distributes cycle counts between = active hyperthreads, i.e., those in C0. A hyperthread becomes inactive whe= n it executes the HLT or MWAIT instructions. If all other hyperthreads are= inactive (or disabled or do not exist), all counts are attributed to this = hyperthread. To obtain the full count when the Core is active, sum the coun= ts from each hyperthread. Available PDIST counters: 0", + "PublicDescription": "This event distributes cycle counts between = active hyperthreads, i.e., those in C0. A hyperthread becomes inactive whe= n it executes the HLT or MWAIT instructions. If all other hyperthreads are= inactive (or disabled or do not exist), all counts are attributed to this = hyperthread. To obtain the full count when the Core is active, sum the coun= ts from each hyperthread.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -307,7 +306,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE", - "PublicDescription": "Counts Core crystal clock cycles when curren= t thread is unhalted and the other thread is halted. Available PDIST counte= rs: 0", + "PublicDescription": "Counts Core crystal clock cycles when curren= t thread is unhalted and the other thread is halted.", "SampleAfterValue": "25003", "UMask": "0x2" }, @@ -316,7 +315,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.PAUSE", - "PublicDescription": "CPU_CLK_UNHALTED.PAUSE Available PDIST count= ers: 0", "SampleAfterValue": "2000003", "UMask": "0x40" }, @@ -327,7 +325,6 @@ "EdgeDetect": "1", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.PAUSE_INST", - "PublicDescription": "CPU_CLK_UNHALTED.PAUSE_INST Available PDIST = counters: 0", "SampleAfterValue": "2000003", "UMask": "0x40" }, @@ -336,7 +333,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.REF_DISTRIBUTED", - "PublicDescription": "This event distributes Core crystal clock cy= cle counts between active hyperthreads, i.e., those in C0 sleep-state. A hy= perthread becomes inactive when it executes the HLT or MWAIT instructions. = If one thread is active in a core, all counts are attributed to this hypert= hread. To obtain the full count when the Core is active, sum the counts fro= m each hyperthread. Available PDIST counters: 0", + "PublicDescription": "This event distributes Core crystal clock cy= cle counts between active hyperthreads, i.e., those in C0 sleep-state. A hy= perthread becomes inactive when it executes the HLT or MWAIT instructions. = If one thread is active in a core, all counts are attributed to this hypert= hread. To obtain the full count when the Core is active, sum the counts fro= m each hyperthread.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -344,7 +341,7 @@ "BriefDescription": "Reference cycles when the core is not in halt= state.", "Counter": "Fixed counter 2", "EventName": "CPU_CLK_UNHALTED.REF_TSC", - "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the eight programmable counte= rs available for other events. Note: On all current platforms this event st= ops counting during 'throttling (TM)' states duty off periods the processor= is 'halted'. The counter update is done at a lower clock rate then the co= re clock the overflow status bit for this counter may appear 'sticky'. Aft= er the counter has overflowed and software clears the overflow status bit a= nd resets the counter to less than MAX. The reset value to the counter is n= ot clocked immediately so the overflow status bit will flip 'high (1)' and = generate another PMI (if enabled) after which the reset value gets clocked = into the counter. Therefore, software will get the interrupt, read the over= flow status bit '1 for bit 34 while the counter value is less than MAX. Sof= tware should ignore this case. Available PDIST counters: 0", + "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the eight programmable counte= rs available for other events. Note: On all current platforms this event st= ops counting during 'throttling (TM)' states duty off periods the processor= is 'halted'. The counter update is done at a lower clock rate then the co= re clock the overflow status bit for this counter may appear 'sticky'. Aft= er the counter has overflowed and software clears the overflow status bit a= nd resets the counter to less than MAX. The reset value to the counter is n= ot clocked immediately so the overflow status bit will flip 'high (1)' and = generate another PMI (if enabled) after which the reset value gets clocked = into the counter. Therefore, software will get the interrupt, read the over= flow status bit '1 for bit 34 while the counter value is less than MAX. Sof= tware should ignore this case.", "SampleAfterValue": "2000003", "UMask": "0x3" }, @@ -353,7 +350,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.REF_TSC_P", - "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the four (eight when Hyperthr= eading is disabled) programmable counters available for other events. Note:= On all current platforms this event stops counting during 'throttling (TM)= ' states duty off periods the processor is 'halted'. The counter update is= done at a lower clock rate then the core clock the overflow status bit for= this counter may appear 'sticky'. After the counter has overflowed and so= ftware clears the overflow status bit and resets the counter to less than M= AX. The reset value to the counter is not clocked immediately so the overfl= ow status bit will flip 'high (1)' and generate another PMI (if enabled) af= ter which the reset value gets clocked into the counter. Therefore, softwar= e will get the interrupt, read the overflow status bit '1 for bit 34 while = the counter value is less than MAX. Software should ignore this case. Avail= able PDIST counters: 0", + "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the four (eight when Hyperthr= eading is disabled) programmable counters available for other events. Note:= On all current platforms this event stops counting during 'throttling (TM)= ' states duty off periods the processor is 'halted'. The counter update is= done at a lower clock rate then the core clock the overflow status bit for= this counter may appear 'sticky'. After the counter has overflowed and so= ftware clears the overflow status bit and resets the counter to less than M= AX. The reset value to the counter is not clocked immediately so the overfl= ow status bit will flip 'high (1)' and generate another PMI (if enabled) af= ter which the reset value gets clocked into the counter. Therefore, softwar= e will get the interrupt, read the overflow status bit '1 for bit 34 while = the counter value is less than MAX. Software should ignore this case.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -361,7 +358,7 @@ "BriefDescription": "Core cycles when the thread is not in halt st= ate", "Counter": "Fixed counter 1", "EventName": "CPU_CLK_UNHALTED.THREAD", - "PublicDescription": "Counts the number of core cycles while the t= hread is not in a halt state. The thread enters the halt state when it is r= unning the HLT instruction. This event is a component in many key event rat= ios. The core frequency may change from time to time due to transitions ass= ociated with Enhanced Intel SpeedStep Technology or TM2. For this reason th= is event may have a changing ratio with regards to time. When the core freq= uency is constant, this event can approximate elapsed time while the core w= as not in the halt state. It is counted on a dedicated fixed counter, leavi= ng the eight programmable counters available for other events. Available PD= IST counters: 0", + "PublicDescription": "Counts the number of core cycles while the t= hread is not in a halt state. The thread enters the halt state when it is r= unning the HLT instruction. This event is a component in many key event rat= ios. The core frequency may change from time to time due to transitions ass= ociated with Enhanced Intel SpeedStep Technology or TM2. For this reason th= is event may have a changing ratio with regards to time. When the core freq= uency is constant, this event can approximate elapsed time while the core w= as not in the halt state. It is counted on a dedicated fixed counter, leavi= ng the eight programmable counters available for other events.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -370,7 +367,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.THREAD_P", - "PublicDescription": "This is an architectural event that counts t= he number of thread cycles while the thread is not in a halt state. The thr= ead enters the halt state when it is running the HLT instruction. The core = frequency may change from time to time due to power or thermal throttling. = For this reason, this event may have a changing ratio with regards to wall = clock time. Available PDIST counters: 0", + "PublicDescription": "This is an architectural event that counts t= he number of thread cycles while the thread is not in a halt state. The thr= ead enters the halt state when it is running the HLT instruction. The core = frequency may change from time to time due to power or thermal throttling. = For this reason, this event may have a changing ratio with regards to wall = clock time.", "SampleAfterValue": "2000003" }, { @@ -379,7 +376,6 @@ "CounterMask": "8", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS", - "PublicDescription": "Cycles while L1 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -389,7 +385,6 @@ "CounterMask": "1", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS", - "PublicDescription": "Cycles while L2 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -399,7 +394,6 @@ "CounterMask": "16", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY", - "PublicDescription": "Cycles while memory subsystem has an outstan= ding load. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -409,7 +403,6 @@ "CounterMask": "12", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS", - "PublicDescription": "Execution stalls while L1 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0xc" }, @@ -419,7 +412,6 @@ "CounterMask": "5", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS", - "PublicDescription": "Execution stalls while L2 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x5" }, @@ -429,7 +421,6 @@ "CounterMask": "4", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL", - "PublicDescription": "Total execution stalls. Available PDIST coun= ters: 0", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -438,7 +429,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb7", "EventName": "EXE.AMX_BUSY", - "PublicDescription": "Counts the cycles where the AMX (Advance Mat= rix Extension) unit is busy performing an operation. Available PDIST counte= rs: 0", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -447,7 +437,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.1_PORTS_UTIL", - "PublicDescription": "Counts cycles during which a total of 1 uop = was executed on all ports and Reservation Station (RS) was not empty. Avail= able PDIST counters: 0", + "PublicDescription": "Counts cycles during which a total of 1 uop = was executed on all ports and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -456,7 +446,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.2_3_PORTS_UTIL", - "PublicDescription": "Cycles total of 2 or 3 uops are executed on = all ports and Reservation Station (RS) was not empty. Available PDIST count= ers: 0", "SampleAfterValue": "2000003", "UMask": "0xc" }, @@ -465,7 +454,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.2_PORTS_UTIL", - "PublicDescription": "Counts cycles during which a total of 2 uops= were executed on all ports and Reservation Station (RS) was not empty. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts cycles during which a total of 2 uops= were executed on all ports and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -474,7 +463,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.3_PORTS_UTIL", - "PublicDescription": "Cycles total of 3 uops are executed on all p= orts and Reservation Station (RS) was not empty. Available PDIST counters: = 0", + "PublicDescription": "Cycles total of 3 uops are executed on all p= orts and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -483,7 +472,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.4_PORTS_UTIL", - "PublicDescription": "Cycles total of 4 uops are executed on all p= orts and Reservation Station (RS) was not empty. Available PDIST counters: = 0", + "PublicDescription": "Cycles total of 4 uops are executed on all p= orts and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -493,7 +482,6 @@ "CounterMask": "5", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.BOUND_ON_LOADS", - "PublicDescription": "Execution stalls while memory subsystem has = an outstanding load. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x21" }, @@ -503,7 +491,7 @@ "CounterMask": "2", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.BOUND_ON_STORES", - "PublicDescription": "Counts cycles where the Store Buffer was ful= l and no loads caused an execution stall. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where the Store Buffer was ful= l and no loads caused an execution stall.", "SampleAfterValue": "1000003", "UMask": "0x40" }, @@ -512,7 +500,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.EXE_BOUND_0_PORTS", - "PublicDescription": "Number of cycles total of 0 uops executed on= all ports, Reservation Station (RS) was not empty, the Store Buffer (SB) w= as not full and there was no outstanding load. Available PDIST counters: 0", + "PublicDescription": "Number of cycles total of 0 uops executed on= all ports, Reservation Station (RS) was not empty, the Store Buffer (SB) w= as not full and there was no outstanding load.", "SampleAfterValue": "1000003", "UMask": "0x80" }, @@ -521,7 +509,7 @@ "Counter": "0,1,2,3", "EventCode": "0x75", "EventName": "INST_DECODED.DECODERS", - "PublicDescription": "Number of decoders utilized in a cycle when = the MITE (legacy decode pipeline) fetches instructions. Available PDIST cou= nters: 0", + "PublicDescription": "Number of decoders utilized in a cycle when = the MITE (legacy decode pipeline) fetches instructions.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -546,7 +534,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.MACRO_FUSED", - "PublicDescription": "INST_RETIRED.MACRO_FUSED Available PDIST cou= nters: 0", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -555,7 +542,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.NOP", - "PublicDescription": "Counts all retired NOP or ENDBR32/64 or PREF= ETCHIT0/1 instructions Available PDIST counters: 0", + "PublicDescription": "Counts all retired NOP or ENDBR32/64 or PREF= ETCHIT0/1 instructions", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -572,7 +559,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.REP_ITERATION", - "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent. Available PDIST counters: 0", + "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -583,7 +570,7 @@ "EdgeDetect": "1", "EventCode": "0xad", "EventName": "INT_MISC.CLEARS_COUNT", - "PublicDescription": "Counts the number of speculative clears due = to any type of branch misprediction or machine clears Available PDIST count= ers: 0", + "PublicDescription": "Counts the number of speculative clears due = to any type of branch misprediction or machine clears", "SampleAfterValue": "500009", "UMask": "0x1" }, @@ -592,7 +579,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.CLEAR_RESTEER_CYCLES", - "PublicDescription": "Cycles after recovery from a branch mispredi= ction or machine clear till the first uop is issued from the resteered path= . Available PDIST counters: 0", + "PublicDescription": "Cycles after recovery from a branch mispredi= ction or machine clear till the first uop is issued from the resteered path= .", "SampleAfterValue": "500009", "UMask": "0x80" }, @@ -601,7 +588,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.MBA_STALLS", - "PublicDescription": "INT_MISC.MBA_STALLS Available PDIST counters= : 0", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -610,7 +596,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.RECOVERY_CYCLES", - "PublicDescription": "Counts core cycles when the Resource allocat= or was stalled due to recovery from an earlier branch misprediction or mach= ine clear event. Available PDIST counters: 0", + "PublicDescription": "Counts core cycles when the Resource allocat= or was stalled due to recovery from an earlier branch misprediction or mach= ine clear event.", "SampleAfterValue": "500009", "UMask": "0x1" }, @@ -621,7 +607,6 @@ "EventName": "INT_MISC.UNKNOWN_BRANCH_CYCLES", "MSRIndex": "0x3F7", "MSRValue": "0x7", - "PublicDescription": "Bubble cycles of BAClear (Unknown Branch). A= vailable PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x40" }, @@ -630,7 +615,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.UOP_DROPPING", - "PublicDescription": "Estimated number of Top-down Microarchitectu= re Analysis slots that got dropped due to non front-end reasons Available P= DIST counters: 0", + "PublicDescription": "Estimated number of Top-down Microarchitectu= re Analysis slots that got dropped due to non front-end reasons", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -639,7 +624,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.128BIT", - "PublicDescription": "INT_VEC_RETIRED.128BIT Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0x13" }, @@ -648,7 +632,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.256BIT", - "PublicDescription": "INT_VEC_RETIRED.256BIT Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0xac" }, @@ -657,7 +640,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.ADD_128", - "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 128-bit vector instructions. Available PDIST counters: 0= ", + "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 128-bit vector instructions.", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -666,7 +649,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.ADD_256", - "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 256-bit vector instructions. Available PDIST counters: 0= ", + "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 256-bit vector instructions.", "SampleAfterValue": "1000003", "UMask": "0xc" }, @@ -675,7 +658,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.MUL_256", - "PublicDescription": "INT_VEC_RETIRED.MUL_256 Available PDIST coun= ters: 0", "SampleAfterValue": "1000003", "UMask": "0x80" }, @@ -684,7 +666,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.SHUFFLES", - "PublicDescription": "INT_VEC_RETIRED.SHUFFLES Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x40" }, @@ -693,7 +674,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.VNNI_128", - "PublicDescription": "INT_VEC_RETIRED.VNNI_128 Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -702,7 +682,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.VNNI_256", - "PublicDescription": "INT_VEC_RETIRED.VNNI_256 Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -711,7 +690,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.ADDRESS_ALIAS", - "PublicDescription": "Counts the number of times a load got blocke= d due to false dependencies in MOB due to partial compare on address. Avail= able PDIST counters: 0", + "PublicDescription": "Counts the number of times a load got blocke= d due to false dependencies in MOB due to partial compare on address.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -720,7 +699,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.NO_SR", - "PublicDescription": "Counts the number of times that split load o= perations are temporarily blocked because all resources for handling the sp= lit accesses are in use. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times that split load o= perations are temporarily blocked because all resources for handling the sp= lit accesses are in use.", "SampleAfterValue": "100003", "UMask": "0x88" }, @@ -729,7 +708,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.STORE_FORWARD", - "PublicDescription": "Counts the number of times where store forwa= rding was prevented for a load operation. The most common case is a load bl= ocked due to the address of memory access (partially) overlapping with a pr= eceding uncompleted store. Note: See the table of not supported store forwa= rds in the Optimization Guide. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times where store forwa= rding was prevented for a load operation. The most common case is a load bl= ocked due to the address of memory access (partially) overlapping with a pr= eceding uncompleted store. Note: See the table of not supported store forwa= rds in the Optimization Guide.", "SampleAfterValue": "100003", "UMask": "0x82" }, @@ -738,7 +717,7 @@ "Counter": "0,1,2,3", "EventCode": "0x4c", "EventName": "LOAD_HIT_PREFETCH.SWPF", - "PublicDescription": "Counts all software-prefetch load dispatches= that hit the fill buffer (FB) allocated for the software prefetch. It can = also be incremented by some lock instructions. So it should only be used wi= th profiling so that the locks can be excluded by ASM (Assembly File) inspe= ction of the nearby instructions. Available PDIST counters: 0", + "PublicDescription": "Counts all software-prefetch load dispatches= that hit the fill buffer (FB) allocated for the software prefetch. It can = also be incremented by some lock instructions. So it should only be used wi= th profiling so that the locks can be excluded by ASM (Assembly File) inspe= ction of the nearby instructions.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -748,7 +727,7 @@ "CounterMask": "1", "EventCode": "0xa8", "EventName": "LSD.CYCLES_ACTIVE", - "PublicDescription": "Counts the cycles when at least one uop is d= elivered by the LSD (Loop-stream detector). Available PDIST counters: 0", + "PublicDescription": "Counts the cycles when at least one uop is d= elivered by the LSD (Loop-stream detector).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -758,7 +737,7 @@ "CounterMask": "6", "EventCode": "0xa8", "EventName": "LSD.CYCLES_OK", - "PublicDescription": "Counts the cycles when optimal number of uop= s is delivered by the LSD (Loop-stream detector). Available PDIST counters:= 0", + "PublicDescription": "Counts the cycles when optimal number of uop= s is delivered by the LSD (Loop-stream detector).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -767,7 +746,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa8", "EventName": "LSD.UOPS", - "PublicDescription": "Counts the number of uops delivered to the b= ack-end by the LSD(Loop Stream Detector). Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops delivered to the b= ack-end by the LSD(Loop Stream Detector).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -778,7 +757,7 @@ "EdgeDetect": "1", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.COUNT", - "PublicDescription": "Counts the number of machine clears (nukes) = of any type. Available PDIST counters: 0", + "PublicDescription": "Counts the number of machine clears (nukes) = of any type.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -787,7 +766,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.SMC", - "PublicDescription": "Counts self-modifying code (SMC) detected, w= hich causes a machine clear. Available PDIST counters: 0", + "PublicDescription": "Counts self-modifying code (SMC) detected, w= hich causes a machine clear.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -796,7 +775,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe0", "EventName": "MISC2_RETIRED.LFENCE", - "PublicDescription": "number of LFENCE retired instructions Availa= ble PDIST counters: 0", + "PublicDescription": "number of LFENCE retired instructions", "SampleAfterValue": "400009", "UMask": "0x20" }, @@ -805,7 +784,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcc", "EventName": "MISC_RETIRED.LBR_INSERTS", - "PublicDescription": "Increments when an entry is added to the Las= t Branch Record (LBR) array (or removed from the array in case of RETURNs i= n call stack mode). The event requires LBR enable via IA32_DEBUGCTL MSR and= branch type selection via MSR_LBR_SELECT. Available PDIST counters: 0", + "PublicDescription": "Increments when an entry is added to the Las= t Branch Record (LBR) array (or removed from the array in case of RETURNs i= n call stack mode). The event requires LBR enable via IA32_DEBUGCTL MSR and= branch type selection via MSR_LBR_SELECT.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -814,7 +793,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa2", "EventName": "RESOURCE_STALLS.SB", - "PublicDescription": "Counts allocation stall cycles caused by the= store buffer (SB) being full. This counts cycles that the pipeline back-en= d blocked uop delivery from the front-end. Available PDIST counters: 0", + "PublicDescription": "Counts allocation stall cycles caused by the= store buffer (SB) being full. This counts cycles that the pipeline back-en= d blocked uop delivery from the front-end.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -823,7 +802,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa2", "EventName": "RESOURCE_STALLS.SCOREBOARD", - "PublicDescription": "Counts cycles where the pipeline is stalled = due to serializing operations. Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -832,7 +810,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa5", "EventName": "RS.EMPTY", - "PublicDescription": "Counts cycles during which the reservation s= tation (RS) is empty for this logical processor. This is usually caused whe= n the front-end pipeline runs into starvation periods (e.g. branch mispredi= ctions or i-cache misses) Available PDIST counters: 0", + "PublicDescription": "Counts cycles during which the reservation s= tation (RS) is empty for this logical processor. This is usually caused whe= n the front-end pipeline runs into starvation periods (e.g. branch mispredi= ctions or i-cache misses)", "SampleAfterValue": "1000003", "UMask": "0x7" }, @@ -844,7 +822,7 @@ "EventCode": "0xa5", "EventName": "RS.EMPTY_COUNT", "Invert": "1", - "PublicDescription": "Counts end of periods where the Reservation = Station (RS) was empty. Could be useful to closely sample on front-end late= ncy issues (see the FRONTEND_RETIRED event of designated precise events) Av= ailable PDIST counters: 0", + "PublicDescription": "Counts end of periods where the Reservation = Station (RS) was empty. Could be useful to closely sample on front-end late= ncy issues (see the FRONTEND_RETIRED event of designated precise events)", "SampleAfterValue": "100003", "UMask": "0x7" }, @@ -853,7 +831,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa5", "EventName": "RS.EMPTY_RESOURCE", - "PublicDescription": "Cycles when RS was empty and a resource allo= cation stall is asserted Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -862,7 +839,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.BACKEND_BOUND_SLOTS", - "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that were not consumed by the back-end pipeline due to lack of ba= ck-end resources, as a result of memory subsystem delays, execution units l= imitations, or other conditions. The count is distributed among unhalted lo= gical processors (hyper-threads) who share the same physical core, in proce= ssors that support Intel Hyper-Threading Technology. Software can use this = event as the numerator for the Backend Bound metric (or top-level category)= of the Top-down Microarchitecture Analysis method. Available PDIST counter= s: 0", + "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that were not consumed by the back-end pipeline due to lack of ba= ck-end resources, as a result of memory subsystem delays, execution units l= imitations, or other conditions. The count is distributed among unhalted lo= gical processors (hyper-threads) who share the same physical core, in proce= ssors that support Intel Hyper-Threading Technology. Software can use this = event as the numerator for the Backend Bound metric (or top-level category)= of the Top-down Microarchitecture Analysis method.", "SampleAfterValue": "10000003", "UMask": "0x2" }, @@ -871,7 +848,7 @@ "Counter": "0", "EventCode": "0xa4", "EventName": "TOPDOWN.BAD_SPEC_SLOTS", - "PublicDescription": "Number of slots of TMA method that were wast= ed due to incorrect speculation. It covers all types of control-flow or dat= a-related mis-speculations. Available PDIST counters: 0", + "PublicDescription": "Number of slots of TMA method that were wast= ed due to incorrect speculation. It covers all types of control-flow or dat= a-related mis-speculations.", "SampleAfterValue": "10000003", "UMask": "0x4" }, @@ -880,7 +857,7 @@ "Counter": "0", "EventCode": "0xa4", "EventName": "TOPDOWN.BR_MISPREDICT_SLOTS", - "PublicDescription": "Number of TMA slots that were wasted due to = incorrect speculation by (any type of) branch mispredictions. This event es= timates number of speculative operations that were issued but not retired a= s well as the out-of-order engine recovery past a branch misprediction. Ava= ilable PDIST counters: 0", + "PublicDescription": "Number of TMA slots that were wasted due to = incorrect speculation by (any type of) branch mispredictions. This event es= timates number of speculative operations that were issued but not retired a= s well as the out-of-order engine recovery past a branch misprediction.", "SampleAfterValue": "10000003", "UMask": "0x8" }, @@ -889,7 +866,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.MEMORY_BOUND_SLOTS", - "PublicDescription": "TOPDOWN.MEMORY_BOUND_SLOTS Available PDIST c= ounters: 0", "SampleAfterValue": "10000003", "UMask": "0x10" }, @@ -897,7 +873,7 @@ "BriefDescription": "TMA slots available for an unhalted logical p= rocessor. Fixed counter - architectural event", "Counter": "Fixed counter 3", "EventName": "TOPDOWN.SLOTS", - "PublicDescription": "Number of available slots for an unhalted lo= gical processor. The event increments by machine-width of the narrowest pip= eline as employed by the Top-down Microarchitecture Analysis method (TMA). = The count is distributed among unhalted logical processors (hyper-threads) = who share the same physical core. Software can use this event as the denomi= nator for the top-level metrics of the TMA method. This architectural event= is counted on a designated fixed counter (Fixed Counter 3). Available PDIS= T counters: 0", + "PublicDescription": "Number of available slots for an unhalted lo= gical processor. The event increments by machine-width of the narrowest pip= eline as employed by the Top-down Microarchitecture Analysis method (TMA). = The count is distributed among unhalted logical processors (hyper-threads) = who share the same physical core. Software can use this event as the denomi= nator for the top-level metrics of the TMA method. This architectural event= is counted on a designated fixed counter (Fixed Counter 3).", "SampleAfterValue": "10000003", "UMask": "0x4" }, @@ -906,7 +882,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.SLOTS_P", - "PublicDescription": "Counts the number of available slots for an = unhalted logical processor. The event increments by machine-width of the na= rrowest pipeline as employed by the Top-down Microarchitecture Analysis met= hod. The count is distributed among unhalted logical processors (hyper-thre= ads) who share the same physical core. Available PDIST counters: 0", + "PublicDescription": "Counts the number of available slots for an = unhalted logical processor. The event increments by machine-width of the na= rrowest pipeline as employed by the Top-down Microarchitecture Analysis met= hod. The count is distributed among unhalted logical processors (hyper-thre= ads) who share the same physical core.", "SampleAfterValue": "10000003", "UMask": "0x1" }, @@ -915,7 +891,7 @@ "Counter": "0,1,2,3", "EventCode": "0x76", "EventName": "UOPS_DECODED.DEC0_UOPS", - "PublicDescription": "This event counts the number of not dec-by-a= ll uops decoded by decoder 0. Available PDIST counters: 0", + "PublicDescription": "This event counts the number of not dec-by-a= ll uops decoded by decoder 0.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -924,7 +900,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_0", - "PublicDescription": "Number of uops dispatch to execution port 0= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 0= .", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -933,7 +909,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_1", - "PublicDescription": "Number of uops dispatch to execution port 1= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 1= .", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -942,7 +918,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_2_3_10", - "PublicDescription": "Number of uops dispatch to execution ports 2= , 3 and 10 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 2= , 3 and 10", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -951,7 +927,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_4_9", - "PublicDescription": "Number of uops dispatch to execution ports 4= and 9 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 4= and 9", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -960,7 +936,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_5_11", - "PublicDescription": "Number of uops dispatch to execution ports 5= and 11 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 5= and 11", "SampleAfterValue": "2000003", "UMask": "0x20" }, @@ -969,7 +945,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_6", - "PublicDescription": "Number of uops dispatch to execution port 6= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 6= .", "SampleAfterValue": "2000003", "UMask": "0x40" }, @@ -978,7 +954,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_7_8", - "PublicDescription": "Number of uops dispatch to execution ports = 7 and 8. Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports = 7 and 8.", "SampleAfterValue": "2000003", "UMask": "0x80" }, @@ -987,7 +963,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE", - "PublicDescription": "Counts the number of uops executed from any = thread. Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops executed from any = thread.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -997,7 +973,7 @@ "CounterMask": "1", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1", - "PublicDescription": "Counts cycles when at least 1 micro-op is ex= ecuted from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 1 micro-op is ex= ecuted from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -1007,7 +983,7 @@ "CounterMask": "2", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2", - "PublicDescription": "Counts cycles when at least 2 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 2 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -1017,7 +993,7 @@ "CounterMask": "3", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3", - "PublicDescription": "Counts cycles when at least 3 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 3 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -1027,7 +1003,7 @@ "CounterMask": "4", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4", - "PublicDescription": "Counts cycles when at least 4 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 4 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -1037,7 +1013,7 @@ "CounterMask": "1", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_1", - "PublicDescription": "Cycles where at least 1 uop was executed per= -thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 1 uop was executed per= -thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1047,7 +1023,7 @@ "CounterMask": "2", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_2", - "PublicDescription": "Cycles where at least 2 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 2 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1057,7 +1033,7 @@ "CounterMask": "3", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_3", - "PublicDescription": "Cycles where at least 3 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 3 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1067,7 +1043,7 @@ "CounterMask": "4", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_4", - "PublicDescription": "Cycles where at least 4 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 4 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1078,7 +1054,7 @@ "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.STALLS", "Invert": "1", - "PublicDescription": "Counts cycles during which no uops were disp= atched from the Reservation Station (RS) per thread. Available PDIST counte= rs: 0", + "PublicDescription": "Counts cycles during which no uops were disp= atched from the Reservation Station (RS) per thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1087,7 +1063,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.THREAD", - "PublicDescription": "Counts the number of uops to be executed per= -thread each cycle. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1096,7 +1071,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.X87", - "PublicDescription": "Counts the number of x87 uops executed. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts the number of x87 uops executed.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -1105,7 +1080,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xae", "EventName": "UOPS_ISSUED.ANY", - "PublicDescription": "Counts the number of uops that the Resource = Allocation Table (RAT) issues to the Reservation Station (RS). Available PD= IST counters: 0", + "PublicDescription": "Counts the number of uops that the Resource = Allocation Table (RAT) issues to the Reservation Station (RS).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1115,7 +1090,6 @@ "CounterMask": "1", "EventCode": "0xae", "EventName": "UOPS_ISSUED.CYCLES", - "PublicDescription": "UOPS_ISSUED.CYCLES Available PDIST counters:= 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1125,7 +1099,7 @@ "CounterMask": "1", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.CYCLES", - "PublicDescription": "Counts cycles where at least one uop has ret= ired. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where at least one uop has ret= ired.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1134,7 +1108,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.HEAVY", - "PublicDescription": "Counts the number of retired micro-operation= s (uops) except the last uop of each instruction. An instruction that is de= coded into less than two uops does not contribute to the count. Available P= DIST counters: 0", + "PublicDescription": "Counts the number of retired micro-operation= s (uops) except the last uop of each instruction. An instruction that is de= coded into less than two uops does not contribute to the count.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1145,7 +1119,6 @@ "EventName": "UOPS_RETIRED.MS", "MSRIndex": "0x3F7", "MSRValue": "0x8", - "PublicDescription": "UOPS_RETIRED.MS Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -1154,7 +1127,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.SLOTS", - "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that are utilized by operations that eventually get retired (comm= itted) by the processor pipeline. Usually, this event positively correlates= with higher performance for example, as measured by the instructions-per-= cycle metric. Software can use this event as the numerator for the Retiring= metric (or top-level category) of the Top-down Microarchitecture Analysis = method. Available PDIST counters: 0", + "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that are utilized by operations that eventually get retired (comm= itted) by the processor pipeline. Usually, this event positively correlates= with higher performance for example, as measured by the instructions-per-= cycle metric. Software can use this event as the numerator for the Retiring= metric (or top-level category) of the Top-down Microarchitecture Analysis = method.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -1165,7 +1138,7 @@ "EventCode": "0xc2", "EventName": "UOPS_RETIRED.STALLS", "Invert": "1", - "PublicDescription": "This event counts cycles without actually re= tired uops. Available PDIST counters: 0", + "PublicDescription": "This event counts cycles without actually re= tired uops.", "SampleAfterValue": "1000003", "UMask": "0x2" } diff --git a/tools/perf/pmu-events/arch/x86/graniterapids/uncore-io.json b/= tools/perf/pmu-events/arch/x86/graniterapids/uncore-io.json index f4f956966e16..2ea2637df3fb 100644 --- a/tools/perf/pmu-events/arch/x86/graniterapids/uncore-io.json +++ b/tools/perf/pmu-events/arch/x86/graniterapids/uncore-io.json @@ -1321,7 +1321,6 @@ "FCMask": "0x01", "PerPkg": "1", "PortMask": "0x0FF", - "PublicDescription": "-", "UMask": "0x4", "Unit": "IIO" }, diff --git a/tools/perf/pmu-events/arch/x86/graniterapids/uncore-memory.jso= n b/tools/perf/pmu-events/arch/x86/graniterapids/uncore-memory.json index b991f6e1afbe..9d385be59e3d 100644 --- a/tools/perf/pmu-events/arch/x86/graniterapids/uncore-memory.json +++ b/tools/perf/pmu-events/arch/x86/graniterapids/uncore-memory.json @@ -195,7 +195,6 @@ "EventName": "UNC_M_MR4_2XREF_CYCLES.SCH0_DIMM0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x1", "Unit": "IMC" }, @@ -206,7 +205,6 @@ "EventName": "UNC_M_MR4_2XREF_CYCLES.SCH0_DIMM1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x2", "Unit": "IMC" }, @@ -217,7 +215,6 @@ "EventName": "UNC_M_MR4_2XREF_CYCLES.SCH1_DIMM0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x4", "Unit": "IMC" }, @@ -228,7 +225,6 @@ "EventName": "UNC_M_MR4_2XREF_CYCLES.SCH1_DIMM1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x8", "Unit": "IMC" }, @@ -239,7 +235,6 @@ "EventName": "UNC_M_PDC_MR4ACTIVE_CYCLES.SCH0_DIMM0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x1", "Unit": "IMC" }, @@ -250,7 +245,6 @@ "EventName": "UNC_M_PDC_MR4ACTIVE_CYCLES.SCH0_DIMM1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x2", "Unit": "IMC" }, @@ -261,7 +255,6 @@ "EventName": "UNC_M_PDC_MR4ACTIVE_CYCLES.SCH1_DIMM0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x4", "Unit": "IMC" }, @@ -272,7 +265,6 @@ "EventName": "UNC_M_PDC_MR4ACTIVE_CYCLES.SCH1_DIMM1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x8", "Unit": "IMC" }, @@ -283,7 +275,6 @@ "EventName": "UNC_M_POWERDOWN_CYCLES.SCH0_RANK0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x1", "Unit": "IMC" }, @@ -294,7 +285,6 @@ "EventName": "UNC_M_POWERDOWN_CYCLES.SCH0_RANK1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x2", "Unit": "IMC" }, @@ -305,7 +295,6 @@ "EventName": "UNC_M_POWERDOWN_CYCLES.SCH0_RANK2", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x4", "Unit": "IMC" }, @@ -316,7 +305,6 @@ "EventName": "UNC_M_POWERDOWN_CYCLES.SCH0_RANK3", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x8", "Unit": "IMC" }, @@ -327,7 +315,6 @@ "EventName": "UNC_M_POWERDOWN_CYCLES.SCH1_RANK0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x10", "Unit": "IMC" }, @@ -338,7 +325,6 @@ "EventName": "UNC_M_POWERDOWN_CYCLES.SCH1_RANK1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x20", "Unit": "IMC" }, @@ -349,7 +335,6 @@ "EventName": "UNC_M_POWERDOWN_CYCLES.SCH1_RANK2", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x40", "Unit": "IMC" }, @@ -360,7 +345,6 @@ "EventName": "UNC_M_POWERDOWN_CYCLES.SCH1_RANK3", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x80", "Unit": "IMC" }, @@ -371,7 +355,6 @@ "EventName": "UNC_M_POWER_CHANNEL_PPD_CYCLES", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "Unit": "IMC" }, { @@ -381,7 +364,6 @@ "EventName": "UNC_M_POWER_CRITICAL_THROTTLE_CYCLES.SLOT0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x1", "Unit": "IMC" }, @@ -392,7 +374,6 @@ "EventName": "UNC_M_POWER_CRITICAL_THROTTLE_CYCLES.SLOT1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x2", "Unit": "IMC" }, @@ -423,7 +404,6 @@ "EventName": "UNC_M_POWER_THROTTLE_CYCLES.MR4BLKEN", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x8", "Unit": "IMC" }, @@ -434,7 +414,6 @@ "EventName": "UNC_M_POWER_THROTTLE_CYCLES.RAPLBLK", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x4", "Unit": "IMC" }, @@ -615,7 +594,6 @@ "EventName": "UNC_M_SELF_REFRESH.ENTER_SUCCESS", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "UNC_M_SELF_REFRESH.ENTER_SUCCESS", "UMask": "0x2", "Unit": "IMC" }, @@ -626,7 +604,6 @@ "EventName": "UNC_M_SELF_REFRESH.ENTER_SUCCESS_CYCLES", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x1", "Unit": "IMC" }, @@ -637,7 +614,6 @@ "EventName": "UNC_M_THROTTLE_CRIT_CYCLES.SLOT0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x1", "Unit": "IMC" }, @@ -648,7 +624,6 @@ "EventName": "UNC_M_THROTTLE_CRIT_CYCLES.SLOT1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x2", "Unit": "IMC" }, @@ -659,7 +634,6 @@ "EventName": "UNC_M_THROTTLE_HIGH_CYCLES.SLOT0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x1", "Unit": "IMC" }, @@ -670,7 +644,6 @@ "EventName": "UNC_M_THROTTLE_HIGH_CYCLES.SLOT1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x2", "Unit": "IMC" }, @@ -681,7 +654,6 @@ "EventName": "UNC_M_THROTTLE_LOW_CYCLES.SLOT0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x1", "Unit": "IMC" }, @@ -692,7 +664,6 @@ "EventName": "UNC_M_THROTTLE_LOW_CYCLES.SLOT1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x2", "Unit": "IMC" }, @@ -703,7 +674,6 @@ "EventName": "UNC_M_THROTTLE_MID_CYCLES.SLOT0", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x1", "Unit": "IMC" }, @@ -714,7 +684,6 @@ "EventName": "UNC_M_THROTTLE_MID_CYCLES.SLOT1", "Experimental": "1", "PerPkg": "1", - "PublicDescription": "-", "UMask": "0x2", "Unit": "IMC" }, diff --git a/tools/perf/pmu-events/arch/x86/graniterapids/virtual-memory.js= on b/tools/perf/pmu-events/arch/x86/graniterapids/virtual-memory.json index 3d3f88600e26..609a9549cbf3 100644 --- a/tools/perf/pmu-events/arch/x86/graniterapids/virtual-memory.json +++ b/tools/perf/pmu-events/arch/x86/graniterapids/virtual-memory.json @@ -4,7 +4,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.STLB_HIT", - "PublicDescription": "Counts loads that miss the DTLB (Data TLB) a= nd hit the STLB (Second level TLB). Available PDIST counters: 0", + "PublicDescription": "Counts loads that miss the DTLB (Data TLB) a= nd hit the STLB (Second level TLB).", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -14,7 +14,7 @@ "CounterMask": "1", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a demand load. Available PDIST cou= nters: 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a demand load.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -23,7 +23,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data loads. This implies it missed in the DTLB and furth= er levels of TLB. The page walk can end with or without a fault. Available = PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data loads. This implies it missed in the DTLB and furth= er levels of TLB. The page walk can end with or without a fault.", "SampleAfterValue": "100003", "UMask": "0xe" }, @@ -32,7 +32,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_1G", - "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -41,7 +41,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data loads. This implies address translations missed in the= DTLB and further levels of TLB. The page walk can end with or without a fa= ult. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data loads. This implies address translations missed in the= DTLB and further levels of TLB. The page walk can end with or without a fa= ult.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -50,7 +50,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -59,7 +59,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for a demand load in the PMH (Page Miss Handler) each cycle. Available PDIS= T counters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for a demand load in the PMH (Page Miss Handler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -68,7 +68,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.STLB_HIT", - "PublicDescription": "Counts stores that miss the DTLB (Data TLB) = and hit the STLB (2nd Level TLB). Available PDIST counters: 0", + "PublicDescription": "Counts stores that miss the DTLB (Data TLB) = and hit the STLB (2nd Level TLB).", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -78,7 +78,7 @@ "CounterMask": "1", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a store. Available PDIST counters:= 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a store.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -87,7 +87,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data stores. This implies it missed in the DTLB and furt= her levels of TLB. The page walk can end with or without a fault. Available= PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data stores. This implies it missed in the DTLB and furt= her levels of TLB. The page walk can end with or without a fault.", "SampleAfterValue": "100003", "UMask": "0xe" }, @@ -96,7 +96,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_1G", - "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -105,7 +105,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data stores. This implies address translations missed in th= e DTLB and further levels of TLB. The page walk can end with or without a f= ault. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data stores. This implies address translations missed in th= e DTLB and further levels of TLB. The page walk can end with or without a f= ault.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -114,7 +114,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -123,7 +123,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for a store in the PMH (Page Miss Handler) each cycle. Available PDIST coun= ters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for a store in the PMH (Page Miss Handler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -132,7 +132,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.STLB_HIT", - "PublicDescription": "Counts instruction fetch requests that miss = the ITLB (Instruction TLB) and hit the STLB (Second-level TLB). Available P= DIST counters: 0", + "PublicDescription": "Counts instruction fetch requests that miss = the ITLB (Instruction TLB) and hit the STLB (Second-level TLB).", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -142,7 +142,7 @@ "CounterMask": "1", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a code (instruction fetch) request= . Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a code (instruction fetch) request= .", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -151,7 +151,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes)= caused by a code fetch. This implies it missed in the ITLB (Instruction TL= B) and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes)= caused by a code fetch. This implies it missed in the ITLB (Instruction TL= B) and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0xe" }, @@ -160,7 +160,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M page size= s) caused by a code fetch. This implies it missed in the ITLB (Instruction = TLB) and further levels of TLB. The page walk can end with or without a fau= lt. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M page size= s) caused by a code fetch. This implies it missed in the ITLB (Instruction = TLB) and further levels of TLB. The page walk can end with or without a fau= lt.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -169,7 +169,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K page sizes) = caused by a code fetch. This implies it missed in the ITLB (Instruction TLB= ) and further levels of TLB. The page walk can end with or without a fault.= Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K page sizes) = caused by a code fetch. This implies it missed in the ITLB (Instruction TLB= ) and further levels of TLB. The page walk can end with or without a fault.= ", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -178,7 +178,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for an outstanding code (instruction fetch) request in the PMH (Page Miss H= andler) each cycle. Available PDIST counters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for an outstanding code (instruction fetch) request in the PMH (Page Miss H= andler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10" } diff --git a/tools/perf/pmu-events/arch/x86/mapfile.csv b/tools/perf/pmu-ev= ents/arch/x86/mapfile.csv index e3ac2dede20a..2d7a9fa055dd 100644 --- a/tools/perf/pmu-events/arch/x86/mapfile.csv +++ b/tools/perf/pmu-events/arch/x86/mapfile.csv @@ -13,7 +13,7 @@ GenuineIntel-6-CF,v1.16,emeraldrapids,core GenuineIntel-6-5[CF],v13,goldmont,core GenuineIntel-6-7A,v1.01,goldmontplus,core GenuineIntel-6-B6,v1.09,grandridge,core -GenuineIntel-6-A[DE],v1.10,graniterapids,core +GenuineIntel-6-A[DE],v1.12,graniterapids,core GenuineIntel-6-(3C|45|46),v36,haswell,core GenuineIntel-6-3F,v29,haswellx,core GenuineIntel-6-7[DE],v1.24,icelake,core --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pg1-f202.google.com (mail-pg1-f202.google.com [209.85.215.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id CB309276031 for ; Mon, 18 Aug 2025 19:05:43 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.215.202 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543946; cv=none; b=H0cQ0vBdmEBAtgPbW+kkCy45ESL1z1KHR/sWP/+j0wnuG+sKTXFPW/92SpwmibTNo6J03TmTiXEM1SVxRNJb8xEisIzAA4HZlCzQZ/95qAjR3VBbXMGiL/KvLvASrgp0Gi1557dOr6GTXQN/MVLJyC743WEmV1m2IiLoQdickvU= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543946; c=relaxed/simple; bh=vWKJBpaScNOWQfCoq/T/wkqQUgU2SHHR+ai5/eBBOxE=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=XdQB4p49kJOjRtg37/IXlyPrdqaeLg7ZOtD/xkps5KeKL50swuxsz2RF+tg1o8XNh2Xqo2Gx3z65MHftof7ceQaIFZjNr5bSAMfhoJSr9WJMMPTzv4+XtKPBI9IyXisekNbTF6rgAv1VF0ljmjEJVRvofCJKbsZegkJuXVt2/7w= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=nj2dFUlm; arc=none smtp.client-ip=209.85.215.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="nj2dFUlm" Received: by mail-pg1-f202.google.com with SMTP id 41be03b00d2f7-b471757dec5so8058522a12.3 for ; Mon, 18 Aug 2025 12:05:43 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543943; x=1756148743; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=h5w4JuxdJH1+BngTPD2SES+TYEhCQ4dCjZ3RtaRCrrY=; b=nj2dFUlmmjWKoV0niK7ot62xk6QwoqxxmpRd41n3sXGqFhk4esbfhNvqpQl+kHaHBJ 0BDj81KOQhyufTHITpYyowkjDcKyVOP34AdOSWWLHDbOW+BLCyMso50l0gq2WtaUWtaw dDGb8Me5PsyNllPZHIEkeUWc2NOFMN0rg7R+Lmxq6w89sLbDNeZ9y52jlKHXlHBdYY7w 2sel55lEzHa9FxQ3EMgsyygiY7Iiah89Xsk9e+RzE6SUX7ATRuraJ3ok27Hj0CnHSgRb bGQFtNMQSnjJoxngYUrOjrWrFhd1RLstV3qON8aq7S05xLJlme55gEaigkhqcsa5Rtsp 0Hjw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543943; x=1756148743; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=h5w4JuxdJH1+BngTPD2SES+TYEhCQ4dCjZ3RtaRCrrY=; b=Vzi7D9Ep39w88ZJ8qZpe18M0H8Q8GrjJolEW/Go/EmOuBbZFONEuW547DgFdIr8fbe /zlikSBrCUfhqT6ojVzR0WFOwNK5wbv7AV2P9DxpE9bAv7hWDOhOoZf8CQOZtO2BZ4M2 lFad8VuEYgOG9AH5lmfqh4TBK3GpJ4I7SRYXDafyLm/Etdf3lClHum81Url/RYVcLI1k qe/RsQkyXf0/iO/3pQNXWOAFNN2U4Z1VnVMEkZEhAHbbvIjva9zlmYw84qYh4dyV7U05 GlNE2dwZclxNKTkUEv9CHB3UByNIh4gcovPIC0zGaTlsB0JxzrxdQCnTOsKOdxu32uTU d/ow== X-Forwarded-Encrypted: i=1; AJvYcCVgSDO+xjuXHYo5eEjvgLKvrl8tE8ftWp6mJNshdAMroG80c74QSpmshIXO7gsk8CmiYKCTl9OMKkdpFqQ=@vger.kernel.org X-Gm-Message-State: AOJu0YwwyBr+mfq2mMxtG+8L567NTWU+ngWw8hk3Sr6cnJsHi0WKMQ/3 ys3GcMCcitShuTSITLecw3v9tvqxVRryZar0G/OT3UB+rHTrm4FPOfKLONRKvCaqa/b0Tve3Kpw Ta5v7LADe5A== X-Google-Smtp-Source: AGHT+IFW4wrrGD2D7qkTClbN2j0h9cYk9vkh2VnUtjH3byXgfJffydCOAxZDlAq7xBNP1jmIlLl/8pfatVhS X-Received: from plho15.prod.google.com ([2002:a17:903:23cf:b0:240:468a:9a20]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:903:1965:b0:243:926:b1f3 with SMTP id d9443c01a7336-2446d7a9610mr208817295ad.24.1755543943322; Mon, 18 Aug 2025 12:05:43 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:05 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-10-irogers@google.com> Subject: [PATCH v2 09/20] perf vendor events: Update haswell metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/haswell/hsw-metrics.json | 32 ++++++++--------- .../arch/x86/haswellx/hsx-metrics.json | 35 +++++++++---------- 2 files changed, 31 insertions(+), 36 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/haswell/hsw-metrics.json b/tool= s/perf/pmu-events/arch/x86/haswell/hsw-metrics.json index b26ea70a3628..aebd82ced1cf 100644 --- a/tools/perf/pmu-events/arch/x86/haswell/hsw-metrics.json +++ b/tools/perf/pmu-events/arch/x86/haswell/hsw-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -80,7 +80,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_0 + UOPS_DISPATCHED_PORT= .PORT_1 + UOPS_DISPATCHED_PORT.PORT_5 + UOPS_DISPATCHED_PORT.PORT_6) / tma_= info_thread_slots", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_alu_op_utilization", @@ -98,7 +97,6 @@ }, { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma= _retiring)", "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", @@ -139,7 +137,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU retired uops originated from CISC (complex instruction set computer) in= struction", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "max(0, tma_microcode_sequencer - tma_assists)", "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_gro= up", "MetricName": "tma_cisc", @@ -509,7 +506,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -521,7 +518,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -696,7 +693,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution port for Load operations", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_2 + UOPS_DISPATCHED_PORT= .PORT_3 + UOPS_DISPATCHED_PORT.PORT_7 - UOPS_DISPATCHED_PORT.PORT_4) / (2 *= tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_load_op_utilization", @@ -746,7 +742,7 @@ { "BriefDescription": "This metric represents fraction of slots the = Memory subsystem within the Backend was a bottleneck", "MetricConstraint": "NO_GROUP_EVENTS", - "MetricExpr": "((min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.STALL= S_LDM_PENDING) + RESOURCE_STALLS.SB) / (min(CPU_CLK_UNHALTED.THREAD, CYCLE_= ACTIVITY.CYCLES_NO_EXECUTE) + (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - (cpu= @UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else cpu@UO= PS_EXECUTED.CORE\\,cmask\\=3D2@)) / 2 - (RS_EVENTS.EMPTY_CYCLES if tma_fetc= h_latency > 0.1 else 0) + RESOURCE_STALLS.SB) if #SMT_on else min(CPU_CLK_U= NHALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUTE) + cpu@UOPS_EXECUTED.CORE\= \,cmask\\=3D1@ - (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_= ipc > 1.8 else cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) - (RS_EVENTS.EMPTY_CY= CLES if tma_fetch_latency > 0.1 else 0) + RESOURCE_STALLS.SB) * tma_backend= _bound", + "MetricExpr": "(min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.STALLS= _LDM_PENDING) + RESOURCE_STALLS.SB) / (min(CPU_CLK_UNHALTED.THREAD, CYCLE_A= CTIVITY.CYCLES_NO_EXECUTE) + (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - (cpu@= UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else cpu@UOP= S_EXECUTED.CORE\\,cmask\\=3D2@)) / 2 - (RS_EVENTS.EMPTY_CYCLES if tma_fetch= _latency > 0.1 else 0) + RESOURCE_STALLS.SB if #SMT_on else min(CPU_CLK_UNH= ALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUTE) + cpu@UOPS_EXECUTED.CORE\\,= cmask\\=3D1@ - (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_ip= c > 1.8 else cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) - (RS_EVENTS.EMPTY_CYCL= ES if tma_fetch_latency > 0.1 else 0) + RESOURCE_STALLS.SB) * tma_backend_b= ound", "MetricGroup": "Backend;TmaL2;TopdownL2;tma_L2_group;tma_backend_b= ound_group", "MetricName": "tma_memory_bound", "MetricThreshold": "tma_memory_bound > 0.2 & tma_backend_bound > 0= .2", @@ -856,7 +852,7 @@ { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", "MetricConstraint": "NO_GROUP_EVENTS", - "MetricExpr": "(min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLES= _NO_EXECUTE) + (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - (cpu@UOPS_EXECUTED.= CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else cpu@UOPS_EXECUTED.COR= E\\,cmask\\=3D2@)) / 2 - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.1= else 0) + RESOURCE_STALLS.SB if #SMT_on else min(CPU_CLK_UNHALTED.THREAD, = CYCLE_ACTIVITY.CYCLES_NO_EXECUTE) + cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ -= (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else c= pu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) - (RS_EVENTS.EMPTY_CYCLES if tma_fetc= h_latency > 0.1 else 0) + RESOURCE_STALLS.SB - RESOURCE_STALLS.SB - min(CPU= _CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.STALLS_LDM_PENDING)) / tma_info_thread= _clks", + "MetricExpr": "((min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLE= S_NO_EXECUTE) + (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - (cpu@UOPS_EXECUTED= .CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else cpu@UOPS_EXECUTED.CO= RE\\,cmask\\=3D2@)) / 2 - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.= 1 else 0) + RESOURCE_STALLS.SB if #SMT_on else min(CPU_CLK_UNHALTED.THREAD,= CYCLE_ACTIVITY.CYCLES_NO_EXECUTE) + cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ = - (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else = cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) - (RS_EVENTS.EMPTY_CYCLES if tma_fet= ch_latency > 0.1 else 0) + RESOURCE_STALLS.SB) - RESOURCE_STALLS.SB - min(C= PU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.STALLS_LDM_PENDING)) / tma_info_thre= ad_clks", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", "MetricThreshold": "tma_ports_utilization > 0.15 & (tma_core_bound= > 0.1 & tma_backend_bound > 0.2)", @@ -865,7 +861,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", - "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else (min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUT= E) - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.1 else 0)) / tma_info= _core_core_clks)", + "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUTE= ) - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.1 else 0)) / tma_info_= core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", "MetricThreshold": "tma_ports_utilized_0 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -874,7 +870,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else (cpu@UOPS_EXECUTED.CORE\\= ,cmask\\=3D1@ - cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) / tma_info_core_core= _clks)", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,= cmask\\=3D1@ - cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) / tma_info_core_core_= clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", "MetricThreshold": "tma_ports_utilized_1 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -883,7 +879,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else (cpu@UOPS_EXECUTED.CORE\\= ,cmask\\=3D2@ - cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@) / tma_info_core_core= _clks)", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,= cmask\\=3D2@ - cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@) / tma_info_core_core_= clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", "MetricThreshold": "tma_ports_utilized_2 > 0.15 & (tma_ports_utili= zation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", diff --git a/tools/perf/pmu-events/arch/x86/haswellx/hsx-metrics.json b/too= ls/perf/pmu-events/arch/x86/haswellx/hsx-metrics.json index 8245a98ad4b9..b8845f8a28b9 100644 --- a/tools/perf/pmu-events/arch/x86/haswellx/hsx-metrics.json +++ b/tools/perf/pmu-events/arch/x86/haswellx/hsx-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -282,7 +282,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_0 + UOPS_DISPATCHED_PORT= .PORT_1 + UOPS_DISPATCHED_PORT.PORT_5 + UOPS_DISPATCHED_PORT.PORT_6) / tma_= info_thread_slots", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_alu_op_utilization", @@ -300,7 +299,6 @@ }, { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma= _retiring)", "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", @@ -341,7 +339,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU retired uops originated from CISC (complex instruction set computer) in= struction", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "max(0, tma_microcode_sequencer - tma_assists)", "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_gro= up", "MetricName": "tma_cisc", @@ -711,7 +708,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -723,7 +720,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -756,6 +753,7 @@ }, { "BriefDescription": "Average number of parallel data read requests= to external memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "UNC_C_TOR_OCCUPANCY.MISS_OPCODE@filter_opc\\=3D0x18= 2@ / UNC_C_TOR_OCCUPANCY.MISS_OPCODE@filter_opc\\=3D0x182\\,thresh\\=3D1@", "MetricGroup": "Mem;MemoryBW;SoC", "MetricName": "tma_info_system_mem_parallel_reads", @@ -918,7 +916,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution port for Load operations", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_2 + UOPS_DISPATCHED_PORT= .PORT_3 + UOPS_DISPATCHED_PORT.PORT_7 - UOPS_DISPATCHED_PORT.PORT_4) / (2 *= tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_load_op_utilization", @@ -928,6 +925,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling loads from local memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "200 * (MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM * (= 1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LOA= D_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_UO= PS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM_= LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE= _DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM + MEM_LOAD_UOPS_L3_MISS_R= ETIRED.REMOTE_FWD))) / tma_info_thread_clks", "MetricGroup": "Server;TopdownL5;tma_L5_group;tma_mem_latency_grou= p", "MetricName": "tma_local_mem", @@ -977,7 +975,7 @@ { "BriefDescription": "This metric represents fraction of slots the = Memory subsystem within the Backend was a bottleneck", "MetricConstraint": "NO_GROUP_EVENTS", - "MetricExpr": "((min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.STALL= S_LDM_PENDING) + RESOURCE_STALLS.SB) / (min(CPU_CLK_UNHALTED.THREAD, CYCLE_= ACTIVITY.CYCLES_NO_EXECUTE) + (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - (cpu= @UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else cpu@UO= PS_EXECUTED.CORE\\,cmask\\=3D2@)) / 2 - (RS_EVENTS.EMPTY_CYCLES if tma_fetc= h_latency > 0.1 else 0) + RESOURCE_STALLS.SB) if #SMT_on else min(CPU_CLK_U= NHALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUTE) + cpu@UOPS_EXECUTED.CORE\= \,cmask\\=3D1@ - (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_= ipc > 1.8 else cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) - (RS_EVENTS.EMPTY_CY= CLES if tma_fetch_latency > 0.1 else 0) + RESOURCE_STALLS.SB) * tma_backend= _bound", + "MetricExpr": "(min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.STALLS= _LDM_PENDING) + RESOURCE_STALLS.SB) / (min(CPU_CLK_UNHALTED.THREAD, CYCLE_A= CTIVITY.CYCLES_NO_EXECUTE) + (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - (cpu@= UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else cpu@UOP= S_EXECUTED.CORE\\,cmask\\=3D2@)) / 2 - (RS_EVENTS.EMPTY_CYCLES if tma_fetch= _latency > 0.1 else 0) + RESOURCE_STALLS.SB if #SMT_on else min(CPU_CLK_UNH= ALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUTE) + cpu@UOPS_EXECUTED.CORE\\,= cmask\\=3D1@ - (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_ip= c > 1.8 else cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) - (RS_EVENTS.EMPTY_CYCL= ES if tma_fetch_latency > 0.1 else 0) + RESOURCE_STALLS.SB) * tma_backend_b= ound", "MetricGroup": "Backend;TmaL2;TopdownL2;tma_L2_group;tma_backend_b= ound_group", "MetricName": "tma_memory_bound", "MetricThreshold": "tma_memory_bound > 0.2 & tma_backend_bound > 0= .2", @@ -1087,7 +1085,7 @@ { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", "MetricConstraint": "NO_GROUP_EVENTS", - "MetricExpr": "(min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLES= _NO_EXECUTE) + (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - (cpu@UOPS_EXECUTED.= CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else cpu@UOPS_EXECUTED.COR= E\\,cmask\\=3D2@)) / 2 - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.1= else 0) + RESOURCE_STALLS.SB if #SMT_on else min(CPU_CLK_UNHALTED.THREAD, = CYCLE_ACTIVITY.CYCLES_NO_EXECUTE) + cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ -= (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else c= pu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) - (RS_EVENTS.EMPTY_CYCLES if tma_fetc= h_latency > 0.1 else 0) + RESOURCE_STALLS.SB - RESOURCE_STALLS.SB - min(CPU= _CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.STALLS_LDM_PENDING)) / tma_info_thread= _clks", + "MetricExpr": "((min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLE= S_NO_EXECUTE) + (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - (cpu@UOPS_EXECUTED= .CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else cpu@UOPS_EXECUTED.CO= RE\\,cmask\\=3D2@)) / 2 - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.= 1 else 0) + RESOURCE_STALLS.SB if #SMT_on else min(CPU_CLK_UNHALTED.THREAD,= CYCLE_ACTIVITY.CYCLES_NO_EXECUTE) + cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ = - (cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@ if tma_info_thread_ipc > 1.8 else = cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) - (RS_EVENTS.EMPTY_CYCLES if tma_fet= ch_latency > 0.1 else 0) + RESOURCE_STALLS.SB) - RESOURCE_STALLS.SB - min(C= PU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.STALLS_LDM_PENDING)) / tma_info_thre= ad_clks", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", "MetricThreshold": "tma_ports_utilization > 0.15 & (tma_core_bound= > 0.1 & tma_backend_bound > 0.2)", @@ -1096,7 +1094,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", - "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else (min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUT= E) - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.1 else 0)) / tma_info= _core_core_clks)", + "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUTE= ) - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.1 else 0)) / tma_info_= core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", "MetricThreshold": "tma_ports_utilized_0 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1105,7 +1103,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else (cpu@UOPS_EXECUTED.CORE\\= ,cmask\\=3D1@ - cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) / tma_info_core_core= _clks)", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,= cmask\\=3D1@ - cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@) / tma_info_core_core_= clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", "MetricThreshold": "tma_ports_utilized_1 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1114,7 +1112,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else (cpu@UOPS_EXECUTED.CORE\\= ,cmask\\=3D2@ - cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@) / tma_info_core_core= _clks)", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else cpu@UOPS_EXECUTED.CORE\\,= cmask\\=3D2@ - cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D3@) / tma_info_core_core_= clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", "MetricThreshold": "tma_ports_utilized_2 > 0.15 & (tma_ports_utili= zation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1141,6 +1139,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling loads from remote memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "310 * (MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_DRAM * = (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LO= AD_UOPS_RETIRED.L3_HIT + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT + MEM_LOAD_U= OPS_L3_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS + MEM= _LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOT= E_DRAM + MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM + MEM_LOAD_UOPS_L3_MISS_= RETIRED.REMOTE_FWD))) / tma_info_thread_clks", "MetricGroup": "Server;Snoop;TopdownL5;tma_L5_group;tma_mem_latenc= y_group", "MetricName": "tma_remote_mem", --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pl1-f201.google.com (mail-pl1-f201.google.com [209.85.214.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 2C1BD277016 for ; Mon, 18 Aug 2025 19:05:47 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543953; cv=none; b=iTN0TKA6NstBNXvX37okmfAlN7NnL9pW3Ruz6aBllwQi+cCrm8FMIbVsN6FmS0pTwuyTbgxaNICgiU23scHYXC5gTJOZVlFhkd5mjqajQOKUR74IfrBHKzQs936FxEYyLX4M1uVIFHmCpgQ6l8TUAgcAwqzBwg/EJnV52Q1nzhY= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543953; c=relaxed/simple; bh=Uw/LS7EtIAgH2h4//+yTqy6QpQ3bcS+Ypb7bJ8D6aEk=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=bdpRndSHPgStRUDpWNiHOwQlfsEYBptYKdvsJG9v6NMS4hml/wD+GXXN9fqVHQpZrEr/CTG4F0UZh/U3MenvVcJWH+dZ9vu3T1H4Jf2Rjr5OsoGwLDKgE53JIbwsKHECGm5iF2uf1G1qvfhANZTCzi32aaYdPxGTVHjkg/Mcd14= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=kDAtjrQi; arc=none smtp.client-ip=209.85.214.201 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="kDAtjrQi" Received: by mail-pl1-f201.google.com with SMTP id d9443c01a7336-24458345f5dso47171325ad.3 for ; Mon, 18 Aug 2025 12:05:47 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543947; x=1756148747; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=SPwbVqBETylK+eIDTzEemKkvu5dR7OZKaiRHgM+HqAk=; b=kDAtjrQivf4yeUXYhwTkJK6ERva1uQu4o7DaFmFprqYpiiOwOqGn2PXFsTkgDC92Nl b4AzzRtjlAiTKk8jFi6NnI7//pkIZwu+jqZwEpWFTNd/llZ73mVOKzpy2zAu08dwJC3B k9K5KEurnzlSQCYvF73GQLCG3Tp0qryTJsoljMratxEhNNzwP235m2q5zWyZ/ro9tV1C ri4xaDTVClOVtcURELn3G7tcanhsPfUrCPfMDry8RexJRYUWlBt320Ldl5SJQKoqd5mM WIz5GW4lvVM7LtdaR3SwkleTaeeJixz2wQwDRiLe2SXfw5+1M/FKJYAZblk43+8UVT+W hXYg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543947; x=1756148747; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=SPwbVqBETylK+eIDTzEemKkvu5dR7OZKaiRHgM+HqAk=; b=GUUv0cgSmRQCVmmsldf27ep4iAEpn3L8rrmvdIAKT71WNfwuzCzX2OWirXRvG/+oqD rCuhxlp8e0FQKkbOT24kGzu+n2uK8L/kDuvkS1B1lb/xDfHU8nV+XXxDBKeWPLS3968D IylKzfUaqs68Uc5SQvXssvDG1wUV3buzmtvIkhAv9Hn6tzeeiBincV4ttpBJMYwFaLuq +AMLsm72T3o6Oc3NUOvHvc4YNYegB1hx7SIkcJQhYWhT31BkqVZob65oZ8xVEYKgJzIR beziZ8gwp57wLxzj7/y3+TNaGu7pumAOP6NF4NatS/EqRtGnEQi5R4hTMvK5PJbE2nO6 aPKw== X-Forwarded-Encrypted: i=1; AJvYcCVJfrAu/3Lrx/1IfnsGP7zZ6QZbwXiOMWxyjDcoGNsgaKv7oiBzuMdJU3mkGm3d6wqectkaUfKqAyZ63Nk=@vger.kernel.org X-Gm-Message-State: AOJu0YxSAm+7ZX2mk7lFeH/eaGA2Yp/u1NoCufB/3FcArbOlwIivCS4U tOaFTzAIltrjW92FQmQE3oKFqJxzHMGXeetGSGygzY2Clc3tNUO8lF/OoUti31Zb6QFKSY+7o/n 3qdqmE6ss+g== X-Google-Smtp-Source: AGHT+IEQ/vEol2CS0xiYYibAbIU0M5yFkqAOdF7adSPyGIxPCOEaMVT7HVatJp0/Q+kbGL+pfTTfRbmyO1ob X-Received: from plbks16.prod.google.com ([2002:a17:903:850:b0:240:670f:ad71]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:902:ef0c:b0:240:44a6:5027 with SMTP id d9443c01a7336-24478e42b8bmr137088465ad.15.1755543947368; Mon, 18 Aug 2025 12:05:47 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:06 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-11-irogers@google.com> Subject: [PATCH v2 10/20] perf vendor events: Update icelake metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/icelake/icl-metrics.json | 96 ++++++----- .../arch/x86/icelakex/icx-metrics.json | 155 +++++++++++++----- 2 files changed, 169 insertions(+), 82 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/icelake/icl-metrics.json b/tool= s/perf/pmu-events/arch/x86/icelake/icl-metrics.json index c5bfdb2f288b..cf9ed3edb694 100644 --- a/tools/perf/pmu-events/arch/x86/icelake/icl-metrics.json +++ b/tools/perf/pmu-events/arch/x86/icelake/icl-metrics.json @@ -1,63 +1,63 @@ [ { "BriefDescription": "C10 residency percent per package", - "MetricExpr": "cstate_pkg@c10\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c10\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C10_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C8 residency percent per package", - "MetricExpr": "cstate_pkg@c8\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c8\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C8_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C9 residency percent per package", - "MetricExpr": "cstate_pkg@c9\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c9\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C9_Pkg_Residency", "ScaleUnit": "100%" @@ -85,7 +85,6 @@ }, { "BriefDescription": "This metric estimates how often memory load a= ccesses were aliased by preceding stores (in program order) with a 4K addre= ss offset", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "LD_BLOCKS_PARTIAL.ADDRESS_ALIAS / tma_info_thread_c= lks", "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", "MetricName": "tma_4k_aliasing", @@ -134,6 +133,7 @@ }, { "BriefDescription": "Total pipeline cost of instruction fetch rela= ted bottlenecks by large code footprint programs (i-side cache; TLB and BTB= misses)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * tma_fetch_latency * (tma_itlb_misses + tma_ic= ache_misses + tma_unknown_branches) / (tma_branch_resteers + tma_dsb_switch= es + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches)", "MetricGroup": "BigFootprint;BvBC;Fed;Frontend;IcMiss;MemoryTLB", "MetricName": "tma_bottleneck_big_code", @@ -147,40 +147,45 @@ "MetricThreshold": "tma_bottleneck_branching_overhead > 5", "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_4k_aliasing + tma_dtlb_load + tma_= fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + = tma_store_fwd_blk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l= 1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_b= lk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + = tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_= 4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma= _lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * = (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_boun= d + tma_store_bound)) * (tma_split_loads / (tma_4k_aliasing + tma_dtlb_load= + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_l= oads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_dra= m_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * = (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency + tma_streaming_stores)) + tma_memory_bound * (tma_stor= e_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tm= a_store_bound)) * (tma_store_latency / (tma_dtlb_store + tma_false_sharing = + tma_split_stores + tma_store_latency + tma_streaming_stores)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + 10 * tma_microcode_seq= uencer * tma_other_mispredicts / tma_branch_mispredicts * tma_mispredicts_r= esteers) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_br= anches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tm= a_itlb_misses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * tma_ms /= (tma_dsb + tma_lsd + tma_mite + tma_ms))) - tma_bottleneck_big_code", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + 10 * tma_microcode_seq= uencer * tma_other_mispredicts / tma_branch_mispredicts * tma_mispredicts_r= esteers) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_br= anches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tm= a_itlb_misses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottleneck_big_= code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20" }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cle= ars_resteers + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_b= ranch_mispredicts * tma_mispredicts_resteers) / (tma_clears_resteers + tma_= mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_= dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switc= hes) + tma_fetch_bandwidth * tma_ms / (tma_dsb + tma_lsd + tma_mite + tma_m= s)) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_mis= predicts * tma_branch_mispredicts + tma_machine_clears * tma_other_nukes / = tma_other_nukes + tma_core_bound * (tma_serializing_operation + tma_core_bo= und * RS_EVENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_utilized_0)= / (tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_= microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer)= * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cle= ars_resteers + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_b= ranch_mispredicts * tma_mispredicts_resteers) / (tma_clears_resteers + tma_= mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_= dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switc= hes) + tma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma= _branch_mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_oth= er_nukes / tma_other_nukes + tma_core_bound * (tma_serializing_operation + = tma_core_bound * RS_EVENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_= utilized_0) / (tma_divider + tma_ports_utilization + tma_serializing_operat= ion) + tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode= _sequencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operation= s)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -188,6 +193,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + = tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, tma_4k_aliasing + tm= a_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + = tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound= / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store= _bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_spli= t_stores + tma_store_latency + tma_streaming_stores)))", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", @@ -196,6 +202,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Synchronization= related bottlenecks (data transfers and coherency updates across processor= s)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_l3_bound / (tma_dram= _bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * (t= ma_contested_accesses + tma_data_sharing) / (tma_contested_accesses + tma_d= ata_sharing + tma_l3_hit_latency + tma_sq_full) + tma_store_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * = tma_false_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency + tma_streaming_stores - tma_store_latency)) + tma_mach= ine_clears * (1 - tma_other_nukes / tma_other_nukes))", "MetricGroup": "BvMS;LockCont;Mem;Offcore;tma_issueSyncxn", "MetricName": "tma_bottleneck_memory_synchronization", @@ -204,6 +211,7 @@ }, { "BriefDescription": "Total pipeline cost of Branch Misprediction r= elated bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (1 - 10 * tma_microcode_sequencer * tma_other= _mispredicts / tma_branch_mispredicts) * (tma_branch_mispredicts + tma_fetc= h_latency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switc= hes + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches))", "MetricGroup": "Bad;BadSpec;BrMispredicts;BvMP;tma_issueBM", "MetricName": "tma_bottleneck_mispredictions", @@ -212,7 +220,8 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -220,6 +229,7 @@ }, { "BriefDescription": "Total pipeline cost of \"useful operations\" = - the portion of Retiring category not covered by Branching_Overhead nor Ir= regular_Overhead.", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_retiring - (BR_INST_RETIRED.ALL_BRANCHES= + 2 * BR_INST_RETIRED.NEAR_CALL + INST_RETIRED.NOP) / tma_info_thread_slot= s - tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_se= quencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "BvUW;Ret", "MetricName": "tma_bottleneck_useful_work", @@ -427,7 +437,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%" }, { @@ -619,6 +629,7 @@ }, { "BriefDescription": "Total pipeline cost of DSB (uop cache) hits -= subset of the Instruction_Fetch_BW Bottleneck", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_frontend_bound * (tma_fetch_bandwidth / = (tma_fetch_bandwidth + tma_fetch_latency)) * (tma_dsb / (tma_dsb + tma_lsd = + tma_mite + tma_ms)))", "MetricGroup": "DSB;Fed;FetchBW;tma_issueFB", "MetricName": "tma_info_botlnk_l2_dsb_bandwidth", @@ -1068,7 +1079,7 @@ "MetricName": "tma_info_memory_tlb_store_stlb_mpki" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (UOPS_EXECUTED.CORE_CYCLES_G= E_1 / 2 if #SMT_on else cpu@UOPS_EXECUTED.THREAD\\,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -1091,6 +1102,12 @@ "MetricGroup": "Fed;FetchBW", "MetricName": "tma_info_pipeline_fetch_mite" }, + { + "BriefDescription": "Average number of uops fetched from MS per cy= cle", + "MetricExpr": "IDQ.MS_UOPS / cpu@IDQ.MS_UOPS\\,cmask\\=3D1@", + "MetricGroup": "Fed;FetchLat;MicroSeq", + "MetricName": "tma_info_pipeline_fetch_ms" + }, { "BriefDescription": "Instructions per a microcode Assist invocatio= n", "MetricExpr": "INST_RETIRED.ANY / ASSISTS.ANY", @@ -1107,7 +1124,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -1119,7 +1136,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -1128,7 +1145,7 @@ "MetricExpr": "64 * (UNC_ARB_TRK_REQUESTS.ALL + UNC_ARB_COH_TRK_RE= QUESTS.ALL) / 1e6 / tma_info_system_time / 1e3", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Giga Floating Point Operations Per Second", @@ -1296,12 +1313,12 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", "MetricExpr": "min(2 * (MEM_INST_RETIRED.ALL_LOADS - MEM_LOAD_RETI= RED.FB_HIT - MEM_LOAD_RETIRED.L1_MISS) * 20 / 100, max(CYCLE_ACTIVITY.CYCLE= S_MEM_ANY - CYCLE_ACTIVITY.CYCLES_L1D_MISS, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%" }, { @@ -1325,7 +1342,6 @@ }, { "BriefDescription": "This metric estimates how often the CPU was s= talled due to loads accesses to L3 cache or contended with a sibling Core", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(CYCLE_ACTIVITY.STALLS_L2_MISS - CYCLE_ACTIVITY.STA= LLS_L3_MISS) / tma_info_thread_clks", "MetricGroup": "CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_gr= oup;tma_memory_bound_group", "MetricName": "tma_l3_bound", @@ -1339,7 +1355,7 @@ "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%" }, { @@ -1445,7 +1461,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%" }, { @@ -1454,7 +1470,7 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%" }, { @@ -1522,7 +1538,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "cpu@IDQ.MS_UOPS\\,cmask\\=3D1@ / tma_info_core_core= _clks / 2", + "MetricExpr": "cpu@IDQ.MS_UOPS\\,cmask\\=3D1@ / tma_info_core_core= _clks / 3.3", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -1656,7 +1672,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_info_= thread_slots", + "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -1693,7 +1709,6 @@ }, { "BriefDescription": "This metric represents rate of split store ac= cesses", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "MEM_INST_RETIRED.SPLIT_STORES / tma_info_core_core_= clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_issueSpSt;tma_store_bou= nd_group", "MetricName": "tma_split_stores", @@ -1707,7 +1722,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%" }, { @@ -1721,7 +1736,6 @@ }, { "BriefDescription": "This metric roughly estimates fraction of cyc= les when the memory subsystem had loads blocked since they could not forwar= d data from earlier (in program order) overlapping stores", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * LD_BLOCKS.STORE_FORWARD / tma_info_thread_clks= ", "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", "MetricName": "tma_store_fwd_blk", diff --git a/tools/perf/pmu-events/arch/x86/icelakex/icx-metrics.json b/too= ls/perf/pmu-events/arch/x86/icelakex/icx-metrics.json index a886a0cfee07..f58eec2a1788 100644 --- a/tools/perf/pmu-events/arch/x86/icelakex/icx-metrics.json +++ b/tools/perf/pmu-events/arch/x86/icelakex/icx-metrics.json @@ -1,28 +1,28 @@ [ { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" @@ -78,6 +78,12 @@ "MetricName": "io_bandwidth_read", "ScaleUnit": "1MB/s" }, + { + "BriefDescription": "Bandwidth of inbound IO reads that are initia= ted by end device controllers that are requesting memory from the CPU and m= iss the L3 cache", + "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_MISS_PCIRDCUR * 64 / 1e6 / d= uration_time", + "MetricName": "io_bandwidth_read_l3_miss", + "ScaleUnit": "1MB/s" + }, { "BriefDescription": "Bandwidth of IO reads that are initiated by e= nd device controllers that are requesting memory from the local CPU socket", "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_PCIRDCUR_LOCAL * 64 / 1e6 / = duration_time", @@ -96,6 +102,12 @@ "MetricName": "io_bandwidth_write", "ScaleUnit": "1MB/s" }, + { + "BriefDescription": "Bandwidth of inbound IO writes that are initi= ated by end device controllers that are writing memory to the CPU", + "MetricExpr": "(UNC_CHA_TOR_INSERTS.IO_MISS_ITOM + UNC_CHA_TOR_INS= ERTS.IO_MISS_ITOMCACHENEAR) * 64 / 1e6 / duration_time", + "MetricName": "io_bandwidth_write_l3_miss", + "ScaleUnit": "1MB/s" + }, { "BriefDescription": "Bandwidth of IO writes that are initiated by = end device controllers that are writing memory to the local CPU socket", "MetricExpr": "(UNC_CHA_TOR_INSERTS.IO_ITOM_LOCAL + UNC_CHA_TOR_IN= SERTS.IO_ITOMCACHENEAR_LOCAL) * 64 / 1e6 / duration_time", @@ -108,6 +120,24 @@ "MetricName": "io_bandwidth_write_remote", "ScaleUnit": "1MB/s" }, + { + "BriefDescription": "The percent of inbound full cache line writes= initiated by IO that miss the L3 cache", + "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_MISS_ITOM / UNC_CHA_TOR_INSE= RTS.IO_ITOM", + "MetricName": "io_full_write_l3_miss", + "ScaleUnit": "100%" + }, + { + "BriefDescription": "The percent of inbound partial writes initiat= ed by IO that miss the L3 cache", + "MetricExpr": "(UNC_CHA_TOR_INSERTS.IO_MISS_ITOMCACHENEAR + UNC_CH= A_TOR_INSERTS.IO_MISS_RFO) / (UNC_CHA_TOR_INSERTS.IO_ITOMCACHENEAR + UNC_CH= A_TOR_INSERTS.IO_RFO)", + "MetricName": "io_partial_write_l3_miss", + "ScaleUnit": "100%" + }, + { + "BriefDescription": "The percent of inbound reads initiated by IO = that miss the L3 cache", + "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_MISS_PCIRDCUR / UNC_CHA_TOR_= INSERTS.IO_PCIRDCUR", + "MetricName": "io_read_l3_miss", + "ScaleUnit": "100%" + }, { "BriefDescription": "Ratio of number of completed page walks (for = 2 megabyte and 4 megabyte page sizes) caused by a code fetch to the total n= umber of completed instructions", "MetricExpr": "ITLB_MISSES.WALK_COMPLETED_2M_4M / INST_RETIRED.ANY= ", @@ -331,7 +361,6 @@ }, { "BriefDescription": "This metric estimates how often memory load a= ccesses were aliased by preceding stores (in program order) with a 4K addre= ss offset", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "LD_BLOCKS_PARTIAL.ADDRESS_ALIAS / tma_info_thread_c= lks", "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", "MetricName": "tma_4k_aliasing", @@ -380,6 +409,7 @@ }, { "BriefDescription": "Total pipeline cost of instruction fetch rela= ted bottlenecks by large code footprint programs (i-side cache; TLB and BTB= misses)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * tma_fetch_latency * (tma_itlb_misses + tma_ic= ache_misses + tma_unknown_branches) / (tma_branch_resteers + tma_dsb_switch= es + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches)", "MetricGroup": "BigFootprint;BvBC;Fed;Frontend;IcMiss;MemoryTLB", "MetricName": "tma_bottleneck_big_code", @@ -393,40 +423,45 @@ "MetricThreshold": "tma_bottleneck_branching_overhead > 5", "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_4k_aliasing + tma_dtlb_load + tma_= fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + = tma_store_fwd_blk)))", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_late= ncy)) + tma_memory_bound * (tma_l3_bound / (tma_cxl_mem_bound + tma_dram_bo= und + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma= _sq_full / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_latency = + tma_sq_full)) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + t= ma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_boun= d)) * (tma_fb_full / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l= 1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_b= lk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l= 1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_b= lk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + = tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_= 4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma= _lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * = (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_boun= d + tma_store_bound)) * (tma_split_loads / (tma_4k_aliasing + tma_dtlb_load= + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_l= oads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_dra= m_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * = (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency + tma_streaming_stores)) + tma_memory_bound * (tma_stor= e_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tm= a_store_bound)) * (tma_store_latency / (tma_dtlb_store + tma_false_sharing = + tma_split_stores + tma_store_latency + tma_streaming_stores)))", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latenc= y)) + tma_memory_bound * (tma_l3_bound / (tma_cxl_mem_bound + tma_dram_boun= d + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l= 3_hit_latency / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_lat= ency + tma_sq_full)) + tma_memory_bound * tma_l2_bound / (tma_cxl_mem_bound= + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_= bound) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + tma_dram_b= ound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tm= a_l1_latency_dependency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + = tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_= fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + tma_dra= m_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * = (tma_lock_latency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1= _latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_bl= k)) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + tma_dram_boun= d + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_s= plit_loads / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latenc= y_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + t= ma_memory_bound * (tma_store_bound / (tma_cxl_mem_bound + tma_dram_bound + = tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_split= _stores / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_stor= e_latency + tma_streaming_stores)) + tma_memory_bound * (tma_store_bound / = (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_= bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store + tma_fals= e_sharing + tma_split_stores + tma_store_latency + tma_streaming_stores)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + 10 * tma_microcode_seq= uencer * tma_other_mispredicts / tma_branch_mispredicts * tma_mispredicts_r= esteers) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_br= anches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tm= a_itlb_misses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * tma_ms /= (tma_dsb + tma_mite + tma_ms))) - tma_bottleneck_big_code", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + 10 * tma_microcode_seq= uencer * tma_other_mispredicts / tma_branch_mispredicts * tma_mispredicts_r= esteers) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_br= anches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tm= a_itlb_misses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottleneck_big_= code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20" }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cle= ars_resteers + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_b= ranch_mispredicts * tma_mispredicts_resteers) / (tma_clears_resteers + tma_= mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_= dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switc= hes) + tma_fetch_bandwidth * tma_ms / (tma_dsb + tma_mite + tma_ms)) + 10 *= tma_microcode_sequencer * tma_other_mispredicts / tma_branch_mispredicts *= tma_branch_mispredicts + tma_machine_clears * tma_other_nukes / tma_other_= nukes + tma_core_bound * (tma_serializing_operation + tma_core_bound * RS_E= VENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_utilized_0) / (tma_di= vider + tma_ports_utilization + tma_serializing_operation) + tma_microcode_= sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) * (tma_as= sists / tma_microcode_sequencer) * tma_heavy_operations)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cle= ars_resteers + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_b= ranch_mispredicts * tma_mispredicts_resteers) / (tma_clears_resteers + tma_= mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_= dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switc= hes) + tma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma= _branch_mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_oth= er_nukes / tma_other_nukes + tma_core_bound * (tma_serializing_operation + = tma_core_bound * RS_EVENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_= utilized_0) / (tma_divider + tma_ports_utilization + tma_serializing_operat= ion) + tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode= _sequencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operation= s)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -434,7 +469,8 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", - "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + = tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, tma_4k_aliasing + tm= a_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + = tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound= / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store= _bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_spli= t_stores + tma_store_latency + tma_streaming_stores)))", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bou= nd + tma_l3_bound + tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, = tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency += tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_boun= d * (tma_store_bound / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound += tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_dtlb_store / (tma_d= tlb_store + tma_false_sharing + tma_split_stores + tma_store_latency + tma_= streaming_stores)))", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", "MetricThreshold": "tma_bottleneck_memory_data_tlbs > 20", @@ -442,7 +478,8 @@ }, { "BriefDescription": "Total pipeline cost of Memory Synchronization= related bottlenecks (data transfers and coherency updates across processor= s)", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * = (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) * tma_remote_cach= e / (tma_local_mem + tma_remote_cache + tma_remote_mem) + tma_l3_bound / (t= ma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_boun= d) * (tma_contested_accesses + tma_data_sharing) / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full) + tma_store_bound / = (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bo= und) * tma_false_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_= stores + tma_store_latency + tma_streaming_stores - tma_store_latency)) + t= ma_machine_clears * (1 - tma_other_nukes / tma_other_nukes))", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound) * (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency= )) * tma_remote_cache / (tma_local_mem + tma_remote_cache + tma_remote_mem)= + tma_l3_bound / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_= l2_bound + tma_l3_bound + tma_store_bound) * (tma_contested_accesses + tma_= data_sharing) / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_lat= ency + tma_sq_full) + tma_store_bound / (tma_cxl_mem_bound + tma_dram_bound= + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * tma_fals= e_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_st= ore_latency + tma_streaming_stores - tma_store_latency)) + tma_machine_clea= rs * (1 - tma_other_nukes / tma_other_nukes))", "MetricGroup": "BvMS;LockCont;Mem;Offcore;tma_issueSyncxn", "MetricName": "tma_bottleneck_memory_synchronization", "MetricThreshold": "tma_bottleneck_memory_synchronization > 10", @@ -450,6 +487,7 @@ }, { "BriefDescription": "Total pipeline cost of Branch Misprediction r= elated bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (1 - 10 * tma_microcode_sequencer * tma_other= _mispredicts / tma_branch_mispredicts) * (tma_branch_mispredicts + tma_fetc= h_latency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switc= hes + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches))", "MetricGroup": "Bad;BadSpec;BrMispredicts;BvMP;tma_issueBM", "MetricName": "tma_bottleneck_mispredictions", @@ -458,7 +496,8 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -466,6 +505,7 @@ }, { "BriefDescription": "Total pipeline cost of \"useful operations\" = - the portion of Retiring category not covered by Branching_Overhead nor Ir= regular_Overhead.", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_retiring - (BR_INST_RETIRED.ALL_BRANCHES= + 2 * BR_INST_RETIRED.NEAR_CALL + INST_RETIRED.NOP) / tma_info_thread_slot= s - tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_se= quencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "BvUW;Ret", "MetricName": "tma_bottleneck_useful_work", @@ -584,6 +624,15 @@ "PublicDescription": "This metric represents fraction of slots whe= re Core non-memory issues were of a bottleneck. Shortage in hardware compu= te resources; or dependencies in software's instructions are both categoriz= ed under Core Bound. Hence it may indicate the machine ran out of an out-of= -order resource; certain execution units are overloaded or dependencies in = program's data- or instruction-flow are limiting the performance (e.g. FP-c= hained long-latency arithmetic operations).", "ScaleUnit": "100%" }, + { + "BriefDescription": "This metric roughly estimates (based on idle = latencies) how often the CPU was stalled on accesses to external CXL Memory= by loads (e.g", + "MetricExpr": "(((1 - ((19 * (MEM_LOAD_L3_MISS_RETIRED.REMOTE_DRAM= * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS)) + 10 * (MEM_LO= AD_L3_MISS_RETIRED.LOCAL_DRAM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RET= IRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REMOTE_FWD * (1 + MEM_LOAD_RETIRED= .FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REMOTE_HITM = * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS))) / (19 * (MEM_L= OAD_L3_MISS_RETIRED.REMOTE_DRAM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_R= ETIRED.L1_MISS)) + 10 * (MEM_LOAD_L3_MISS_RETIRED.LOCAL_DRAM * (1 + MEM_LOA= D_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REM= OTE_FWD * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LO= AD_L3_MISS_RETIRED.REMOTE_HITM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RE= TIRED.L1_MISS)) + (25 * (MEM_LOAD_RETIRED.LOCAL_PMM * (1 + MEM_LOAD_RETIRED= .FB_HIT / MEM_LOAD_RETIRED.L1_MISS) if #has_pmem > 0 else 0) + 33 * (MEM_LO= AD_L3_MISS_RETIRED.REMOTE_PMM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RET= IRED.L1_MISS) if #has_pmem > 0 else 0))) if #has_pmem > 0 else 1)) * (CYCLE= _ACTIVITY.STALLS_L3_MISS / tma_info_thread_clks + (CYCLE_ACTIVITY.STALLS_L1= D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma_info_thread_clks - tma_l2_bou= nd) if 1e6 * (MEM_LOAD_L3_MISS_RETIRED.REMOTE_PMM + MEM_LOAD_RETIRED.LOCAL_= PMM) > MEM_LOAD_RETIRED.L1_MISS else 0) if #has_pmem > 0 else 0)", + "MetricGroup": "MemoryBound;Server;TmaL3mem;TopdownL3;tma_L3_group= ;tma_memory_bound_group", + "MetricName": "tma_cxl_mem_bound", + "MetricThreshold": "tma_cxl_mem_bound > 0.1 & (tma_memory_bound > = 0.2 & tma_backend_bound > 0.2)", + "PublicDescription": "This metric roughly estimates (based on idle= latencies) how often the CPU was stalled on accesses to external CXL Memor= y by loads (e.g. 3D-Xpoint (Crystal Ridge, a.k.a. IXP) memory, PMM - Persis= tent Memory Module [from CLX to SPR] or any other CXL Type3 Memory [EMR onw= ards]).", + "ScaleUnit": "100%" + }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to data-sharing ac= cesses", "MetricConstraint": "NO_GROUP_EVENTS", @@ -615,7 +664,7 @@ { "BriefDescription": "This metric estimates how often the CPU was s= talled on accesses to external memory (DRAM) by loads", "MetricConstraint": "NO_GROUP_EVENTS", - "MetricExpr": "CYCLE_ACTIVITY.STALLS_L3_MISS / tma_info_thread_clk= s + (CYCLE_ACTIVITY.STALLS_L1D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma_= info_thread_clks - tma_l2_bound", + "MetricExpr": "(CYCLE_ACTIVITY.STALLS_L3_MISS / tma_info_thread_cl= ks + (CYCLE_ACTIVITY.STALLS_L1D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma= _info_thread_clks - tma_l2_bound - tma_cxl_mem_bound if #has_pmem > 0 else = CYCLE_ACTIVITY.STALLS_L3_MISS / tma_info_thread_clks + (CYCLE_ACTIVITY.STAL= LS_L1D_MISS - CYCLE_ACTIVITY.STALLS_L2_MISS) / tma_info_thread_clks - tma_l= 2_bound)", "MetricGroup": "MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_me= mory_bound_group", "MetricName": "tma_dram_bound", "MetricThreshold": "tma_dram_bound > 0.1 & (tma_memory_bound > 0.2= & tma_backend_bound > 0.2)", @@ -673,7 +722,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%" }, { @@ -865,6 +914,7 @@ }, { "BriefDescription": "Total pipeline cost of DSB (uop cache) hits -= subset of the Instruction_Fetch_BW Bottleneck", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_frontend_bound * (tma_fetch_bandwidth / = (tma_fetch_bandwidth + tma_fetch_latency)) * (tma_dsb / (tma_dsb + tma_mite= + tma_ms)))", "MetricGroup": "DSB;Fed;FetchBW;tma_issueFB", "MetricName": "tma_info_botlnk_l2_dsb_bandwidth", @@ -1320,7 +1370,7 @@ "MetricName": "tma_info_memory_tlb_store_stlb_mpki" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (UOPS_EXECUTED.CORE_CYCLES_G= E_1 / 2 if #SMT_on else cpu@UOPS_EXECUTED.THREAD\\,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -1337,6 +1387,12 @@ "MetricGroup": "Fed;FetchBW", "MetricName": "tma_info_pipeline_fetch_mite" }, + { + "BriefDescription": "Average number of uops fetched from MS per cy= cle", + "MetricExpr": "IDQ.MS_UOPS / cpu@IDQ.MS_UOPS\\,cmask\\=3D1@", + "MetricGroup": "Fed;FetchLat;MicroSeq", + "MetricName": "tma_info_pipeline_fetch_ms" + }, { "BriefDescription": "Instructions per a microcode Assist invocatio= n", "MetricExpr": "INST_RETIRED.ANY / ASSISTS.ANY", @@ -1353,7 +1409,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -1365,16 +1421,28 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, + { + "BriefDescription": "Average 3DXP Memory Bandwidth Use for reads [= GB / sec]", + "MetricExpr": "(64 * UNC_M_PMM_RPQ_INSERTS / 1e9 / tma_info_system= _time if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryBW;Server;SoC", + "MetricName": "tma_info_system_cxl_mem_read_bw" + }, + { + "BriefDescription": "Average 3DXP Memory Bandwidth Use for Writes = [GB / sec]", + "MetricExpr": "(64 * UNC_M_PMM_WPQ_INSERTS / 1e9 / tma_info_system= _time if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryBW;Server;SoC", + "MetricName": "tma_info_system_cxl_mem_write_bw" + }, { "BriefDescription": "Average external Memory Bandwidth Use for rea= ds and writes [GB / sec]", "MetricExpr": "64 * (UNC_M_CAS_COUNT.RD + UNC_M_CAS_COUNT.WR) / 1e= 9 / tma_info_system_time", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Giga Floating Point Operations Per Second", @@ -1433,11 +1501,19 @@ }, { "BriefDescription": "Average number of parallel data read requests= to external memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "UNC_CHA_TOR_OCCUPANCY.IA_MISS_DRD / UNC_CHA_TOR_OCC= UPANCY.IA_MISS_DRD@thresh\\=3D1@", "MetricGroup": "Mem;MemoryBW;SoC", "MetricName": "tma_info_system_mem_parallel_reads", "PublicDescription": "Average number of parallel data read request= s to external memory. Accounts for demand loads and L1/L2 prefetches" }, + { + "BriefDescription": "Average latency of data read request to exter= nal 3D X-Point memory [in nanoseconds]", + "MetricExpr": "(1e9 * (UNC_CHA_TOR_OCCUPANCY.IA_MISS_DRD_PMM / UNC= _CHA_TOR_INSERTS.IA_MISS_DRD_PMM) / cha_0@event\\=3D0x0@ if #has_pmem > 0 e= lse 0)", + "MetricGroup": "MemOffcore;MemoryLat;Server;SoC", + "MetricName": "tma_info_system_mem_pmm_read_latency", + "PublicDescription": "Average latency of data read request to exte= rnal 3D X-Point memory [in nanoseconds]. Accounts for demand loads and L1/L= 2 data-read prefetches" + }, { "BriefDescription": "Average latency of data read request to exter= nal memory (in nanoseconds)", "MetricExpr": "1e9 * (UNC_CHA_TOR_OCCUPANCY.IA_MISS_DRD / UNC_CHA_= TOR_INSERTS.IA_MISS_DRD) / (tma_info_system_socket_clks / tma_info_system_t= ime)", @@ -1590,12 +1666,12 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", "MetricExpr": "min(2 * (MEM_INST_RETIRED.ALL_LOADS - MEM_LOAD_RETI= RED.FB_HIT - MEM_LOAD_RETIRED.L1_MISS) * 20 / 100, max(CYCLE_ACTIVITY.CYCLE= S_MEM_ANY - CYCLE_ACTIVITY.CYCLES_L1D_MISS, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%" }, { @@ -1619,7 +1695,6 @@ }, { "BriefDescription": "This metric estimates how often the CPU was s= talled due to loads accesses to L3 cache or contended with a sibling Core", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(CYCLE_ACTIVITY.STALLS_L2_MISS - CYCLE_ACTIVITY.STA= LLS_L3_MISS) / tma_info_thread_clks", "MetricGroup": "CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_gr= oup;tma_memory_bound_group", "MetricName": "tma_l3_bound", @@ -1633,7 +1708,7 @@ "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%" }, { @@ -1739,7 +1814,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%" }, { @@ -1748,7 +1823,7 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%" }, { @@ -1816,7 +1891,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "cpu@IDQ.MS_UOPS\\,cmask\\=3D1@ / tma_info_core_core= _clks / 2", + "MetricExpr": "cpu@IDQ.MS_UOPS\\,cmask\\=3D1@ / tma_info_core_core= _clks / 3.3", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -1968,7 +2043,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_info_= thread_slots", + "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -2005,7 +2080,6 @@ }, { "BriefDescription": "This metric represents rate of split store ac= cesses", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "MEM_INST_RETIRED.SPLIT_STORES / tma_info_core_core_= clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_issueSpSt;tma_store_bou= nd_group", "MetricName": "tma_split_stores", @@ -2019,7 +2093,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%" }, { @@ -2033,7 +2107,6 @@ }, { "BriefDescription": "This metric roughly estimates fraction of cyc= les when the memory subsystem had loads blocked since they could not forwar= d data from earlier (in program order) overlapping stores", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * LD_BLOCKS.STORE_FORWARD / tma_info_thread_clks= ", "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", "MetricName": "tma_store_fwd_blk", --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pl1-f201.google.com (mail-pl1-f201.google.com [209.85.214.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 46189277C88 for ; Mon, 18 Aug 2025 19:05:53 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543955; cv=none; b=qIMvu0cGOV6l6TzS1CTG2gAsTAsHn8VmSMPCKo88WktDaLDP8bpbs3T2HkEceg0taitnmhPCjDSqPxmPCPd6aswwEq6LluGBHzoucM+QJH8HvxxSqlfMNH19M8PUbeiNCv3vw9Jto99K2clLLYFYsY2Hint48gojLlOsbiaw+WI= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543955; c=relaxed/simple; bh=226ZCX0rwUGU2rjbBeWjW9HBlYWrYpvNLHHvtA09cNo=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=r85M0X5Q3zAFdVon26QwvEMAdCLT/feaWFgbhgU1Y8W8mPbPdUrYC0GcKJwrEbg73MJtygbpddr7YT8YIkNIx+Xpd3urBxVc6ECX6+HNwMuEBt2LrOa+gSU4/6AGpsp26RhW1OjnbfmNpwATGS2UwH0y2fIuqZ8KXjYVxM6e2yc= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=jVEJaJUd; arc=none smtp.client-ip=209.85.214.201 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="jVEJaJUd" Received: by mail-pl1-f201.google.com with SMTP id d9443c01a7336-2445806eab4so45399385ad.1 for ; Mon, 18 Aug 2025 12:05:53 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543953; x=1756148753; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=sgrL13sD4bHYKRGxftI04WUTsTC1o/9zfZyLctrJOZI=; b=jVEJaJUdyVEcEoYxuwr74tCs+jN82XiHbsToFJTzZf5pJGr4oUAJMRoTiENZL54Hi3 BRtgIkreFb5r12+hvq2IJFw3lHiL+CrA7XGzyraDMwmUpHUZ0qPNfPy4ndVsyRf3cfRD BUsVx+Sml3O1WLzYIM6wxjW2bbdGFM4BOLkIREKQxIIKQ/65AdAVaFVLUN17Ay76OcAv 8cFN34ORDF3WkhG+xvw/sI4iy0shELB91CZuvjBlvjN/O3qgP3ZEkZHog8iIfVlv+IOi jlE4I++wMmW0/MZLLiot05OyqXJcbhNCC4AzIk1TGmM950fq3pfdNxb3ba6K7fUZ20uY wHxA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543953; x=1756148753; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=sgrL13sD4bHYKRGxftI04WUTsTC1o/9zfZyLctrJOZI=; b=vvDQsILtL96ONcmERQTxrt+lsbyleFXeUZHdZa9VpRNDtPpG3CT92fSMT7ceZl258C H10GIu58cCcNrjwIsZnfhz32366ACXyklFWRDKFE4TxtGflwvLvEb2MVFxp+7vgxhioW qTrmBZWTGXYRRUBhDBsljDEQuHdwiI/ZYfo7x2UZHwT7weGMSJ9qzwnujfd6QhlXfmoU dozkZ8RsWH5+GkyLaLujOIDXd2czeTfIo0zpjBe7DCumiFtySWD0bKRXOZSoM34rls69 5IKH1TKH8f5VeQkbBXGiL/fgcJdca5hF5J/OnxbBAkz5HvXG+J5Mho5TEebH2ZAo9QGv BKdA== X-Forwarded-Encrypted: i=1; AJvYcCU1aATOG1ORWoSwecZjZAlzRtEe9Xh8uCV/CzN0kWv+aOYKHtrD4+tHR/spqP0jBjYYifU811+PalHL9LY=@vger.kernel.org X-Gm-Message-State: AOJu0YxlnB7COmYCfiCsomW1t58M8ygW3kglEdIOSddWELUv01aK00M5 XTiAq663kaIals7Ysr/viB3rKJSlE4AbFHOiS2c6uLjCEv+4WhdMmueQyOzcOp2jFVUEmGPwCE4 4qz2Nmfv6xw== X-Google-Smtp-Source: AGHT+IESBlPntiJiF/Qmc1tviH8dEytxctqkt2G8Nmk+IUrjU9YNJSlggCpahSR0jB1HwZIPMdEkCTRN1Dq6 X-Received: from plrj13.prod.google.com ([2002:a17:903:28d:b0:242:abd5:b3bf]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:903:2f4d:b0:240:640a:c57b with SMTP id d9443c01a7336-2446d8b4127mr189206115ad.37.1755543952640; Mon, 18 Aug 2025 12:05:52 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:07 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-12-irogers@google.com> Subject: [PATCH v2 11/20] perf vendor events: Update ivybridge/ivytown metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/ivybridge/ivb-metrics.json | 30 ++++++++--------- .../arch/x86/ivytown/ivt-metrics.json | 33 +++++++++---------- 2 files changed, 29 insertions(+), 34 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/ivybridge/ivb-metrics.json b/to= ols/perf/pmu-events/arch/x86/ivybridge/ivb-metrics.json index de651ff9f846..969cb519eec1 100644 --- a/tools/perf/pmu-events/arch/x86/ivybridge/ivb-metrics.json +++ b/tools/perf/pmu-events/arch/x86/ivybridge/ivb-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -80,7 +80,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_0 + UOPS_DISPATCHED_PORT= .PORT_1 + UOPS_DISPATCHED_PORT.PORT_5) / (3 * tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_alu_op_utilization", @@ -98,7 +97,6 @@ }, { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma= _retiring)", "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", @@ -139,7 +137,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU retired uops originated from CISC (complex instruction set computer) in= struction", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "max(0, tma_microcode_sequencer - tma_assists)", "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_gro= up", "MetricName": "tma_cisc", @@ -561,7 +558,7 @@ "MetricThreshold": "tma_info_memory_tlb_page_walks_utilization > 0= .5" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (cpu@UOPS_EXECUTED.CORE\\,cm= ask\\=3D1@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -574,7 +571,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -586,7 +583,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -775,7 +772,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution port for Load operations", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_2 + UOPS_DISPATCHED_PORT= .PORT_3 - UOPS_DISPATCHED_PORT.PORT_4) / (2 * tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_load_op_utilization", @@ -926,7 +922,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", - "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else (min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUT= E) - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.1 else 0)) / tma_info= _core_core_clks)", + "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUTE= ) - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.1 else 0)) / tma_info_= core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", "MetricThreshold": "tma_ports_utilized_0 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -935,7 +931,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else (UOPS_EXECUTED.CYCLES_GE_= 1_UOP_EXEC - UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC) / tma_info_core_core_clks= )", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1= _UOP_EXEC - UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC) / tma_info_core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", "MetricThreshold": "tma_ports_utilized_1 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -944,7 +940,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else (UOPS_EXECUTED.CYCLES_GE_= 2_UOPS_EXEC - UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clk= s)", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_2= _UOPS_EXEC - UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clks= ", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", "MetricThreshold": "tma_ports_utilized_2 > 0.15 & (tma_ports_utili= zation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", diff --git a/tools/perf/pmu-events/arch/x86/ivytown/ivt-metrics.json b/tool= s/perf/pmu-events/arch/x86/ivytown/ivt-metrics.json index 714d5e6d21e7..1cdd197ac883 100644 --- a/tools/perf/pmu-events/arch/x86/ivytown/ivt-metrics.json +++ b/tools/perf/pmu-events/arch/x86/ivytown/ivt-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -80,7 +80,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_0 + UOPS_DISPATCHED_PORT= .PORT_1 + UOPS_DISPATCHED_PORT.PORT_5) / (3 * tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_alu_op_utilization", @@ -98,7 +97,6 @@ }, { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma= _retiring)", "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", @@ -139,7 +137,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU retired uops originated from CISC (complex instruction set computer) in= struction", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "max(0, tma_microcode_sequencer - tma_assists)", "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_gro= up", "MetricName": "tma_cisc", @@ -561,7 +558,7 @@ "MetricThreshold": "tma_info_memory_tlb_page_walks_utilization > 0= .5" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (cpu@UOPS_EXECUTED.CORE\\,cm= ask\\=3D1@ / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1_UOP_EXEC)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -574,7 +571,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -586,7 +583,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -626,6 +623,7 @@ }, { "BriefDescription": "Average number of parallel data read requests= to external memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "UNC_C_TOR_OCCUPANCY.MISS_OPCODE@filter_opc\\=3D0x18= 2@ / UNC_C_TOR_OCCUPANCY.MISS_OPCODE@filter_opc\\=3D0x182\\,thresh\\=3D1@", "MetricGroup": "Mem;MemoryBW;SoC", "MetricName": "tma_info_system_mem_parallel_reads", @@ -795,7 +793,6 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution port for Load operations", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_2 + UOPS_DISPATCHED_PORT= .PORT_3 - UOPS_DISPATCHED_PORT.PORT_4) / (2 * tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_load_op_utilization", @@ -805,6 +802,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling loads from local memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "200 * (MEM_LOAD_UOPS_LLC_MISS_RETIRED.LOCAL_DRAM * = (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_LO= AD_UOPS_RETIRED.LLC_HIT + MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_HIT + MEM_LOAD= _UOPS_LLC_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_MISS += MEM_LOAD_UOPS_LLC_MISS_RETIRED.LOCAL_DRAM + MEM_LOAD_UOPS_LLC_MISS_RETIRED= .REMOTE_DRAM + MEM_LOAD_UOPS_LLC_MISS_RETIRED.REMOTE_HITM + MEM_LOAD_UOPS_L= LC_MISS_RETIRED.REMOTE_FWD))) / tma_info_thread_clks", "MetricGroup": "Server;TopdownL5;tma_L5_group;tma_mem_latency_grou= p", "MetricName": "tma_local_mem", @@ -955,7 +953,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", - "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else (min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUT= E) - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.1 else 0)) / tma_info= _core_core_clks)", + "MetricExpr": "(cpu@UOPS_EXECUTED.CORE\\,inv\\,cmask\\=3D1@ / 2 if= #SMT_on else min(CPU_CLK_UNHALTED.THREAD, CYCLE_ACTIVITY.CYCLES_NO_EXECUTE= ) - (RS_EVENTS.EMPTY_CYCLES if tma_fetch_latency > 0.1 else 0)) / tma_info_= core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", "MetricThreshold": "tma_ports_utilized_0 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -964,7 +962,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else (UOPS_EXECUTED.CYCLES_GE_= 1_UOP_EXEC - UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC) / tma_info_core_core_clks= )", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D1@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D2@) / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_1= _UOP_EXEC - UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC) / tma_info_core_core_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", "MetricThreshold": "tma_ports_utilized_1 > 0.2 & (tma_ports_utiliz= ation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -973,7 +971,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else (UOPS_EXECUTED.CYCLES_GE_= 2_UOPS_EXEC - UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clk= s)", + "MetricExpr": "((cpu@UOPS_EXECUTED.CORE\\,cmask\\=3D2@ - cpu@UOPS_= EXECUTED.CORE\\,cmask\\=3D3@) / 2 if #SMT_on else UOPS_EXECUTED.CYCLES_GE_2= _UOPS_EXEC - UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC) / tma_info_core_core_clks= ", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", "MetricThreshold": "tma_ports_utilized_2 > 0.15 & (tma_ports_utili= zation > 0.15 & (tma_core_bound > 0.1 & tma_backend_bound > 0.2))", @@ -1000,6 +998,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling loads from remote memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "310 * (MEM_LOAD_UOPS_LLC_MISS_RETIRED.REMOTE_DRAM *= (1 + MEM_LOAD_UOPS_RETIRED.HIT_LFB / (MEM_LOAD_UOPS_RETIRED.L2_HIT + MEM_L= OAD_UOPS_RETIRED.LLC_HIT + MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_HIT + MEM_LOA= D_UOPS_LLC_HIT_RETIRED.XSNP_HITM + MEM_LOAD_UOPS_LLC_HIT_RETIRED.XSNP_MISS = + MEM_LOAD_UOPS_LLC_MISS_RETIRED.LOCAL_DRAM + MEM_LOAD_UOPS_LLC_MISS_RETIRE= D.REMOTE_DRAM + MEM_LOAD_UOPS_LLC_MISS_RETIRED.REMOTE_HITM + MEM_LOAD_UOPS_= LLC_MISS_RETIRED.REMOTE_FWD))) / tma_info_thread_clks", "MetricGroup": "Server;Snoop;TopdownL5;tma_L5_group;tma_mem_latenc= y_group", "MetricName": "tma_remote_mem", --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pl1-f201.google.com (mail-pl1-f201.google.com [209.85.214.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 10558274649 for ; Mon, 18 Aug 2025 19:05:58 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543960; cv=none; b=C9heaf/KWMt1wHSI3XnU5sXFBixk1jAdx3L10M8s8iWEtiOc7dIlRAsRfT8nn0ET7L0gqUtBO7J2vbThKbgH4NowefHjlbq4C2KBqzCbHXcB0VeBoAXDA/0hH+SxreKWH+NMufl2hRxejDHkQ4zyQSWqMd7sdGscMzmQw+v8oJA= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543960; c=relaxed/simple; bh=ya0dbwWUzKMKLeRGa0TbpgvQFOApNQzPsgfDgaJe3jo=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=Kvwmo+K3+Q4h+/VQJMl+0o/UBqBANqTCTmZxk420KliS4a28MJc2okRbWgSSQVk9hCOSNLxIy5X/JdihklFCEzeKYZZ4oj3i5sulV1q7IvlKgE4N0AYubCPiajLCOETFgrS+TTr4OLoNzs3eujaPc0zflX/Q6GjyPvDqcXzpNBY= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=3zx6YJqK; arc=none smtp.client-ip=209.85.214.201 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="3zx6YJqK" Received: by mail-pl1-f201.google.com with SMTP id d9443c01a7336-2445806b2beso110478885ad.1 for ; Mon, 18 Aug 2025 12:05:58 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543958; x=1756148758; darn=vger.kernel.org; h=to:from:subject:message-id:references:mime-version:in-reply-to:date :from:to:cc:subject:date:message-id:reply-to; bh=jPO7XFOOsTvMdi14sVo1ekX03/m/Zq/cV95OnqVa63U=; b=3zx6YJqKOkUTOhs7rrUy7Z0gBURcLp08Xco62y7Hvb9J0Z5Si0gD7M4DJqtUEEctBs YD4TWxMK9gFKVvGhKe1upPlCR5BXhlYiZ93vmFkPuiL6obq1BQHkk+M+c5xIqC7sYPkm on9aH54ua1BEK6nKYSXAssuSC7ll8Yte4RbWxkXoslCCvA+Mp9dsgpoUFBoOLcq9jGBC RH9NN52nJ83WTmc/O0RqMVAnb2pf6n0UV8h5se1EMk9eidKyn9OHbPxXOnnggykL/HhG T58/j5GWhY3MQx6oVqwlr8xsXZxwCloKG7qvWczaQabSrZqSBsn5WbXzxLGbwhKEVyvs UhsQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543958; x=1756148758; h=to:from:subject:message-id:references:mime-version:in-reply-to:date :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=jPO7XFOOsTvMdi14sVo1ekX03/m/Zq/cV95OnqVa63U=; b=rCN8HEYvlwh+AQ1GgdWAvDO+Q+QmZVL97pl9u9gCE/MJfvMNwqBREld6ZJhsxKYa+O 03ME+rGHgmQpIhqh3GR9aNIeGvuN6RHJX0glkM5m4F3NDC6njWg63+p1xBFnkESaF6h1 NDxFdfezdIxfQNm2oTBzs+3q3Rj74HATJGG1VU4fn760y/5w/ZeL1LOsBaMZliqpQy2y f3A0dREgiTFsDZRLJzqA3aIp0zRh7JPMe+U2IzwJaWFNNGYlfF+dA9mHNK9ntOLfkak4 MGZBejvKJj4mc6iSsN550cGHA+OcEXDx2Zk0k024o2LmPs8Lg9Uu0DNQzsRm0JE0pTbj wVPw== X-Forwarded-Encrypted: i=1; AJvYcCVTXdVELSGH/FT64Z2h4OvdPax3BK5HlSo9BtJPlRPp5uTneBtk+LckvToeFmNG+Q8qPJ6U6yKqwrOhtiw=@vger.kernel.org X-Gm-Message-State: AOJu0YxzO6U8fHch2BuZhz5rdHwGIuoRDOAQPR7u3hOpJW6e2KxlDuzF NBN79vXlfyiqtS/tzvdhFQxsZceEsE0zXKlkYR3UBNlJYOfhB2bbqOuFuZXKGBNak6A+OQ7H2Bc QpczNm0979g== X-Google-Smtp-Source: AGHT+IEHy4TfrMkq0HMb600V8hJmkl3do3SI7TVe489UJoALKFOP4WLU7LMYMAz+LuWy0oi1SnPpLCv7A7z4 X-Received: from pjc13.prod.google.com ([2002:a17:90b:2f4d:b0:312:14e5:174b]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:902:ced2:b0:242:cf0b:66cd with SMTP id d9443c01a7336-2446d8f0a20mr204219665ad.34.1755543958231; Mon, 18 Aug 2025 12:05:58 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:08 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-13-irogers@google.com> Subject: [PATCH v2 12/20] perf vendor events: Update jaketown metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/jaketown/jkt-metrics.json | 20 +++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/jaketown/jkt-metrics.json b/too= ls/perf/pmu-events/arch/x86/jaketown/jkt-metrics.json index 6f636ea0f216..250c73b21385 100644 --- a/tools/perf/pmu-events/arch/x86/jaketown/jkt-metrics.json +++ b/tools/perf/pmu-events/arch/x86/jaketown/jkt-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -71,7 +71,6 @@ }, { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma= _retiring)", "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", @@ -296,7 +295,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -308,7 +307,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -348,6 +347,7 @@ }, { "BriefDescription": "Average number of parallel data read requests= to external memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "UNC_C_TOR_OCCUPANCY.MISS_OPCODE@filter_opc\\=3D0x18= 2@ / UNC_C_TOR_OCCUPANCY.MISS_OPCODE@filter_opc\\=3D0x182\\,thresh\\=3D1@", "MetricGroup": "Mem;MemoryBW;SoC", "MetricName": "tma_info_system_mem_parallel_reads", --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pf1-f201.google.com (mail-pf1-f201.google.com [209.85.210.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 6A44E311972 for ; Mon, 18 Aug 2025 19:06:04 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.210.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543971; cv=none; b=hra6eD6OAbAfDVzF6dQ1B1XwyTSjZH8vheG8FTfwqM+uuj+aLGv9F1OG4jUUrL7DBvjeKoHfEMtPzFypViFfpEEiSlblQadjMI/aRO/55A6pI5eAf3vxprXc4wyOQMUnXUfJTuzSLFc08VHh8YA4Ed739vnIkAWo2rJRRA1b614= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543971; c=relaxed/simple; bh=PZ6QGv5glp8mZV55Me3zzjdx99SQyXVIkI4dqTighkk=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=p2/MiRNYOj3Hyx16+i9BI8ILC+lxKfxgdOBDvsRp5kRtHTb5MWcxYjbZ0VHq0F2rob3VLHwjzFyRZsrOuGDzKDydIV4CMqgYb3hE5qdTFhZKnKIRV0LCU4ZZtx4TLTKiWdkxttKpJtxfty7M0yvTSQgmnDyveitHLxNf0jFMQfg= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=EweTTYIV; arc=none smtp.client-ip=209.85.210.201 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="EweTTYIV" Received: by mail-pf1-f201.google.com with SMTP id d2e1a72fcca58-76e2e591846so3963263b3a.0 for ; Mon, 18 Aug 2025 12:06:04 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543964; x=1756148764; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=PQKoSQvn8vnhZLYKmQap+v16zDFMThiS9xe5u+Yfpmg=; b=EweTTYIVFlHuVAjmQr49rYkDXTmm0kYJ9TWUIcvQvLaIIWGEIPeSzxIi8FNzvBw9R/ F5wazImIZWbgy/MixHrSWyQchzCEkRIA8qZVVQBiUOcKXEVp3UJgd4g6TXB7PwT+OCnm o09zPt2QYYiOmbJtc1XjuPhBv3GP3NhfjCiYPfbZeLwp0m0Acj0dLZk0KC4nWyOs70B/ Tzq01OaBM6/Tho1v55JX9XPzg1b38HaZ3opwL0ij3nxQ+WYLzfIN7M5yZsrYDiiNsX9o /5wmjMkz0wd2TEiqLTUmCmLq4I7wFBrcjs1B9uD5e7yeLhTgdUMNoJPE04SQJqzqng75 HeGw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543964; x=1756148764; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=PQKoSQvn8vnhZLYKmQap+v16zDFMThiS9xe5u+Yfpmg=; b=nbfLvDtJeb1wfnDRVru2rIt7S5QooF43zBsHQdFaA1Ufzt3z/D2NpgHgENQYuoeKoA Np2yfYR4kR91t9avgV2aY/q02TEzgGEj4G3b4waqrzPIA1mWLGESzT7TDkoDar1jde5/ 7VoXngDDKURDa9fDrBz/5Ky8pH9UV0mu26TW5mmHvaq7j81M5iBEiYqZycb9cpLNk1tD 78wm8pASS2WcuHwuS6m6Jicwjba4MwzF2voXyMSlSkc02+QwIrxF8wvwSS9PMtgh/0J3 7grAD3RJJhuHhc6oG/DLBN7jjgqfxBdURiPvq6ueFiTX5rleK+85+d+J7PbVmJ+vDagx EOsQ== X-Forwarded-Encrypted: i=1; AJvYcCV4Uc4cjv2fdRDSy3R3CPSd18+E47f9QuMBG/1rLxkHVIErcqYRe2njvPrIhHzRZQgdckyY+FdE1LR6bNc=@vger.kernel.org X-Gm-Message-State: AOJu0YwcYUwSQeb9oIlwpS9ceKCRJ90RHeJeNi6675EJTZjZoj4l/Xen UbCKOHFgnhvoIH/jSSRI1SJw4saJeRywk1zEmJp+24W2LfsZ6vz6Su7UtBP3ueRCBxkV7urzsfg ut6I51uAfdA== X-Google-Smtp-Source: AGHT+IHVWV6BtaciNYgDVycp+J/BCDZSjk4rtNglYOnLjBwLo/JsIjVbLUMTmSrHVyQXm+RLN+Gz0q/XQ+w0 X-Received: from pfuv15.prod.google.com ([2002:a05:6a00:148f:b0:76e:3404:b5ca]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a05:6a00:1e13:b0:748:e2d8:100d with SMTP id d2e1a72fcca58-76e7cfc1f87mr552857b3a.8.1755543963445; Mon, 18 Aug 2025 12:06:03 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:09 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-14-irogers@google.com> Subject: [PATCH v2 13/20] perf vendor events: Update lunarlake events/metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update events from v1.14 to v1.17. Update metrics from TMA 5.0 to 5.1. The event updates come from: https://github.com/intel/perfmon/commit/6bdcbce3e9df30ae02bd0ea51fd73bf51ee= 8aff4 https://github.com/intel/perfmon/commit/1684fa543fd45970759bb72dc3fc00c2ef8= 7c0e8 Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../pmu-events/arch/x86/lunarlake/cache.json | 104 ++++++--- .../arch/x86/lunarlake/frontend.json | 40 ++-- .../arch/x86/lunarlake/lnl-metrics.json | 216 ++++++++++-------- .../pmu-events/arch/x86/lunarlake/memory.json | 22 +- .../arch/x86/lunarlake/pipeline.json | 85 ++++--- .../x86/lunarlake/uncore-interconnect.json | 10 + .../arch/x86/lunarlake/uncore-memory.json | 8 + tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- 8 files changed, 294 insertions(+), 193 deletions(-) create mode 100644 tools/perf/pmu-events/arch/x86/lunarlake/uncore-interco= nnect.json diff --git a/tools/perf/pmu-events/arch/x86/lunarlake/cache.json b/tools/pe= rf/pmu-events/arch/x86/lunarlake/cache.json index ff37d49611c3..29bcb847178f 100644 --- a/tools/perf/pmu-events/arch/x86/lunarlake/cache.json +++ b/tools/perf/pmu-events/arch/x86/lunarlake/cache.json @@ -28,6 +28,16 @@ "UMask": "0x1", "Unit": "cpu_core" }, + { + "BriefDescription": "Cachelines replaced into the L1 d-cache. Succ= essful replacements only (not blocked) and exclude WB-miss case", + "Counter": "0,1,2,3,4,5,6,7,8,9", + "EventCode": "0x51", + "EventName": "L1D.L1_REPLACEMENT", + "PublicDescription": "Counts cachelines replaced into the L1 d-cac= he.", + "SampleAfterValue": "1000003", + "UMask": "0x4", + "Unit": "cpu_core" + }, { "BriefDescription": "Cachelines replaced into the L0 and L1 d-cach= e. Successful replacements only (not blocked) and exclude WB-miss case", "Counter": "0,1,2,3,4,5,6,7,8,9", @@ -592,7 +602,7 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.ALL_LOADS", - "PublicDescription": "Counts Instructions with at least one archit= ecturally visible load retired. Available PDIST counters: 0", + "PublicDescription": "Counts Instructions with at least one archit= ecturally visible load retired. Available PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x81", "Unit": "cpu_core" @@ -603,7 +613,7 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.ALL_STORES", - "PublicDescription": "Counts all retired store instructions. Avail= able PDIST counters: 0", + "PublicDescription": "Counts all retired store instructions. Avail= able PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x82", "Unit": "cpu_core" @@ -613,7 +623,7 @@ "Counter": "0,1,2,3", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.ALL_SWPF", - "PublicDescription": "Counts all retired software prefetch instruc= tions. Available PDIST counters: 0", + "PublicDescription": "Counts all retired software prefetch instruc= tions. Available PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x84", "Unit": "cpu_core" @@ -624,7 +634,7 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.ANY", - "PublicDescription": "Counts all retired memory instructions - loa= ds and stores. Available PDIST counters: 0", + "PublicDescription": "Counts all retired memory instructions - loa= ds and stores. Available PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x87", "Unit": "cpu_core" @@ -635,7 +645,7 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.LOCK_LOADS", - "PublicDescription": "Counts retired load instructions with locked= access. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with locked= access. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x21", "Unit": "cpu_core" @@ -646,7 +656,7 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.SPLIT_LOADS", - "PublicDescription": "Counts retired load instructions that split = across a cacheline boundary. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions that split = across a cacheline boundary. Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x41", "Unit": "cpu_core" @@ -657,18 +667,29 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.SPLIT_STORES", - "PublicDescription": "Counts retired store instructions that split= across a cacheline boundary. Available PDIST counters: 0", + "PublicDescription": "Counts retired store instructions that split= across a cacheline boundary. Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x42", "Unit": "cpu_core" }, + { + "BriefDescription": "Retired instructions that hit the STLB.", + "Counter": "0,1,2,3", + "Data_LA": "1", + "EventCode": "0xd0", + "EventName": "MEM_INST_RETIRED.STLB_HIT_ANY", + "PublicDescription": "Number of retired instructions with a clean = hit in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", + "SampleAfterValue": "100003", + "UMask": "0xf", + "Unit": "cpu_core" + }, { "BriefDescription": "Retired load instructions that hit the STLB.", "Counter": "0,1,2,3", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.STLB_HIT_LOADS", - "PublicDescription": "Number of retired load instructions with a c= lean hit in the 2nd-level TLB (STLB). Available PDIST counters: 0", + "PublicDescription": "Number of retired load instructions with a c= lean hit in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x9", "Unit": "cpu_core" @@ -679,18 +700,39 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.STLB_HIT_STORES", - "PublicDescription": "Number of retired store instructions that hi= t in the 2nd-level TLB (STLB). Available PDIST counters: 0", + "PublicDescription": "Number of retired store instructions that hi= t in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0xa", "Unit": "cpu_core" }, + { + "BriefDescription": "Retired SWPF instructions that hit the STLB.", + "Counter": "0,1,2,3", + "EventCode": "0xd0", + "EventName": "MEM_INST_RETIRED.STLB_HIT_SWPF", + "PublicDescription": "Number of retired SWPF instructions that hit= in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", + "SampleAfterValue": "1000003", + "UMask": "0xc", + "Unit": "cpu_core" + }, + { + "BriefDescription": "Retired instructions that miss the STLB.", + "Counter": "0,1,2,3", + "Data_LA": "1", + "EventCode": "0xd0", + "EventName": "MEM_INST_RETIRED.STLB_MISS_ANY", + "PublicDescription": "Retired instructions that miss the STLB. Ava= ilable PDIST counters: 0,1", + "SampleAfterValue": "100003", + "UMask": "0x17", + "Unit": "cpu_core" + }, { "BriefDescription": "Retired load instructions that miss the STLB.= ", "Counter": "0,1,2,3", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.STLB_MISS_LOADS", - "PublicDescription": "Number of retired load instructions that (st= art a) miss in the 2nd-level TLB (STLB). Available PDIST counters: 0", + "PublicDescription": "Number of retired load instructions that (st= art a) miss in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x11", "Unit": "cpu_core" @@ -701,18 +743,28 @@ "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_INST_RETIRED.STLB_MISS_STORES", - "PublicDescription": "Number of retired store instructions that (s= tart a) miss in the 2nd-level TLB (STLB). Available PDIST counters: 0", + "PublicDescription": "Number of retired store instructions that (s= tart a) miss in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x12", "Unit": "cpu_core" }, + { + "BriefDescription": "Retired SWPF instructions that miss the STLB.= ", + "Counter": "0,1,2,3", + "EventCode": "0xd0", + "EventName": "MEM_INST_RETIRED.STLB_MISS_SWPF", + "PublicDescription": "Number of retired SWPF instructions that (st= art a) miss in the 2nd-level TLB (STLB). Available PDIST counters: 0,1", + "SampleAfterValue": "1000003", + "UMask": "0x14", + "Unit": "cpu_core" + }, { "BriefDescription": "Retired load instructions whose data sources = were a cross-core Snoop hits and forwards data from an in on-package core c= ache (induced by NI$)", "Counter": "0,1,2,3", "Data_LA": "1", "EventCode": "0xd2", "EventName": "MEM_LOAD_L3_HIT_RETIRED.XSNP_FWD", - "PublicDescription": "Counts retired load instructions whose data = sources were a cross-core Snoop hits and forwards data from an in on-packag= e core cache (induced by NI$) Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions whose data = sources were a cross-core Snoop hits and forwards data from an in on-packag= e core cache (induced by NI$) Available PDIST counters: 0,1", "SampleAfterValue": "20011", "UMask": "0x10", "Unit": "cpu_core" @@ -723,7 +775,7 @@ "Data_LA": "1", "EventCode": "0xd2", "EventName": "MEM_LOAD_L3_HIT_RETIRED.XSNP_HITM", - "PublicDescription": "Counts retired load instructions whose data = sources were HitM responses from shared L3, Hit-with-FWD is normally exclud= ed. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions whose data = sources were HitM responses from shared L3, Hit-with-FWD is normally exclud= ed. Available PDIST counters: 0,1", "SampleAfterValue": "20011", "UMask": "0x4", "Unit": "cpu_core" @@ -734,7 +786,7 @@ "Data_LA": "1", "EventCode": "0xd2", "EventName": "MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS", - "PublicDescription": "Counts the retired load instructions whose d= ata sources were L3 hit and cross-core snoop missed in on-pkg core cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts the retired load instructions whose d= ata sources were L3 hit and cross-core snoop missed in on-pkg core cache. A= vailable PDIST counters: 0,1", "SampleAfterValue": "20011", "UMask": "0x1", "Unit": "cpu_core" @@ -745,7 +797,7 @@ "Data_LA": "1", "EventCode": "0xd2", "EventName": "MEM_LOAD_L3_HIT_RETIRED.XSNP_NO_FWD", - "PublicDescription": "Counts retired load instructions whose data = sources were L3 and cross-core snoop hits in on-pkg core cache. Available P= DIST counters: 0", + "PublicDescription": "Counts retired load instructions whose data = sources were L3 and cross-core snoop hits in on-pkg core cache. Available P= DIST counters: 0,1", "SampleAfterValue": "20011", "UMask": "0x2", "Unit": "cpu_core" @@ -756,7 +808,7 @@ "Data_LA": "1", "EventCode": "0xd3", "EventName": "MEM_LOAD_L3_MISS_RETIRED.MEMSIDE_CACHE", - "PublicDescription": "Retired load instructions which data source = is memory side cache. Available PDIST counters: 0", + "PublicDescription": "Retired load instructions which data source = is memory side cache. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "Unit": "cpu_core" }, @@ -766,7 +818,7 @@ "Data_LA": "1", "EventCode": "0xd4", "EventName": "MEM_LOAD_MISC_RETIRED.UC", - "PublicDescription": "Retired instructions with at least one load = to uncacheable memory-type, or at least one cache-line split locked access = (Bus Lock). Available PDIST counters: 0", + "PublicDescription": "Retired instructions with at least one load = to uncacheable memory-type, or at least one cache-line split locked access = (Bus Lock). Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x4", "Unit": "cpu_core" @@ -777,7 +829,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.FB_HIT", - "PublicDescription": "Counts retired load instructions with at lea= st one uop was load missed in L1 but hit FB (Fill Buffers) due to preceding= miss to the same cache line with data not ready. Available PDIST counters:= 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop was load missed in L1 but hit FB (Fill Buffers) due to preceding= miss to the same cache line with data not ready. Available PDIST counters:= 0,1", "SampleAfterValue": "100007", "UMask": "0x40", "Unit": "cpu_core" @@ -788,7 +840,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L1_HIT", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the L1 data cache. This event includes all SW prefet= ches and lock instructions regardless of the data source. Available PDIST c= ounters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the L1 data cache. This event includes all SW prefet= ches and lock instructions regardless of the data source. Available PDIST c= ounters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x101", "Unit": "cpu_core" @@ -799,7 +851,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L1_HIT_L0", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the Level 0 of the L1 data cache. This event include= s all SW prefetches and lock instructions regardless of the data source. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the Level 0 of the L1 data cache. This event include= s all SW prefetches and lock instructions regardless of the data source. Av= ailable PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -809,7 +861,7 @@ "Counter": "0,1,2,3", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L1_HIT_L1", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the Level 1 of the L1 data cache. Available PDIST co= unters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the Level 1 of the L1 data cache. Available PDIST co= unters: 0,1", "SampleAfterValue": "1000003", "Unit": "cpu_core" }, @@ -819,7 +871,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L1_MISS", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that missed in the L1 cache. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that missed in the L1 cache. Available PDIST counters: 0,1", "SampleAfterValue": "200003", "UMask": "0x8", "Unit": "cpu_core" @@ -830,7 +882,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L2_HIT", - "PublicDescription": "Counts retired load instructions with L2 cac= he hits as data sources. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with L2 cac= he hits as data sources. Available PDIST counters: 0,1", "SampleAfterValue": "200003", "UMask": "0x2", "Unit": "cpu_core" @@ -841,7 +893,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L2_MISS", - "PublicDescription": "Counts retired load instructions missed L2 c= ache as data sources. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions missed L2 c= ache as data sources. Available PDIST counters: 0,1", "SampleAfterValue": "100021", "UMask": "0x10", "Unit": "cpu_core" @@ -852,7 +904,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L3_HIT", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the L3 cache. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that hit in the L3 cache. Available PDIST counters: 0,1", "SampleAfterValue": "100021", "UMask": "0x4", "Unit": "cpu_core" @@ -863,7 +915,7 @@ "Data_LA": "1", "EventCode": "0xd1", "EventName": "MEM_LOAD_RETIRED.L3_MISS", - "PublicDescription": "Counts retired load instructions with at lea= st one uop that missed in the L3 cache. Available PDIST counters: 0", + "PublicDescription": "Counts retired load instructions with at lea= st one uop that missed in the L3 cache. Available PDIST counters: 0,1", "SampleAfterValue": "50021", "UMask": "0x20", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/lunarlake/frontend.json b/tools= /perf/pmu-events/arch/x86/lunarlake/frontend.json index e2facc4086e9..b21d602e9f1a 100644 --- a/tools/perf/pmu-events/arch/x86/lunarlake/frontend.json +++ b/tools/perf/pmu-events/arch/x86/lunarlake/frontend.json @@ -108,7 +108,7 @@ "EventName": "FRONTEND_RETIRED.ANY_ANT", "MSRIndex": "0x3F7", "MSRValue": "0x9", - "PublicDescription": "Always Not Taken (ANT) conditional retired b= ranches (no BTB entry and not mispredicted) Available PDIST counters: 0", + "PublicDescription": "Always Not Taken (ANT) conditional retired b= ranches (no BTB entry and not mispredicted) Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -120,7 +120,7 @@ "EventName": "FRONTEND_RETIRED.ANY_DSB_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x1", - "PublicDescription": "Counts retired Instructions that experienced= DSB (Decode stream buffer i.e. the decoded instruction-cache) miss. Availa= ble PDIST counters: 0", + "PublicDescription": "Counts retired Instructions that experienced= DSB (Decode stream buffer i.e. the decoded instruction-cache) miss. Availa= ble PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -169,7 +169,7 @@ "EventName": "FRONTEND_RETIRED.DSB_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x11", - "PublicDescription": "Number of retired Instructions that experien= ced a critical DSB (Decode stream buffer i.e. the decoded instruction-cache= ) miss. Critical means stalls were exposed to the back-end as a result of t= he DSB miss. Available PDIST counters: 0", + "PublicDescription": "Number of retired Instructions that experien= ced a critical DSB (Decode stream buffer i.e. the decoded instruction-cache= ) miss. Critical means stalls were exposed to the back-end as a result of t= he DSB miss. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -199,7 +199,7 @@ "EventName": "FRONTEND_RETIRED.ITLB_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x14", - "PublicDescription": "Counts retired Instructions that experienced= iTLB (Instruction TLB) true miss. Available PDIST counters: 0", + "PublicDescription": "Counts retired Instructions that experienced= iTLB (Instruction TLB) true miss. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -211,7 +211,7 @@ "EventName": "FRONTEND_RETIRED.L1I_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x12", - "PublicDescription": "Counts retired Instructions who experienced = Instruction L1 Cache true miss. Available PDIST counters: 0", + "PublicDescription": "Counts retired Instructions who experienced = Instruction L1 Cache true miss. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -223,7 +223,7 @@ "EventName": "FRONTEND_RETIRED.L2_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x13", - "PublicDescription": "Counts retired Instructions who experienced = Instruction L2 Cache true miss. Available PDIST counters: 0", + "PublicDescription": "Counts retired Instructions who experienced = Instruction L2 Cache true miss. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -235,7 +235,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_128", "MSRIndex": "0x3F7", "MSRValue": "0x608006", - "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 12= 8 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0", + "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 12= 8 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -247,7 +247,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_16", "MSRIndex": "0x3F7", "MSRValue": "0x601006", - "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 16 cycles. During th= is period the front-end delivered no uops. Available PDIST counters: 0", + "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 16 cycles. During th= is period the front-end delivered no uops. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -259,7 +259,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_2", "MSRIndex": "0x3F7", "MSRValue": "0x600206", - "PublicDescription": "Retired instructions that are fetched after = an interval where the front-end delivered no uops for a period of at least = 2 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0", + "PublicDescription": "Retired instructions that are fetched after = an interval where the front-end delivered no uops for a period of at least = 2 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -271,7 +271,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_256", "MSRIndex": "0x3F7", "MSRValue": "0x610006", - "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 25= 6 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0", + "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 25= 6 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -283,7 +283,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_2_BUBBLES_GE_1", "MSRIndex": "0x3F7", "MSRValue": "0x100206", - "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after the front-end had at least 1 bubble-slot for a per= iod of 2 cycles. A bubble-slot is an empty issue-pipeline slot while there = was no RAT stall. Available PDIST counters: 0", + "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after the front-end had at least 1 bubble-slot for a per= iod of 2 cycles. A bubble-slot is an empty issue-pipeline slot while there = was no RAT stall. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -295,7 +295,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_32", "MSRIndex": "0x3F7", "MSRValue": "0x602006", - "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 32 cycles. During th= is period the front-end delivered no uops. Available PDIST counters: 0", + "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 32 cycles. During th= is period the front-end delivered no uops. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -307,7 +307,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_4", "MSRIndex": "0x3F7", "MSRValue": "0x600406", - "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 4 = cycles which was not interrupted by a back-end stall. Available PDIST count= ers: 0", + "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 4 = cycles which was not interrupted by a back-end stall. Available PDIST count= ers: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -319,7 +319,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_512", "MSRIndex": "0x3F7", "MSRValue": "0x620006", - "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 51= 2 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0", + "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 51= 2 cycles which was not interrupted by a back-end stall. Available PDIST cou= nters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -331,7 +331,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_64", "MSRIndex": "0x3F7", "MSRValue": "0x604006", - "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 64= cycles which was not interrupted by a back-end stall. Available PDIST coun= ters: 0", + "PublicDescription": "Counts retired instructions that are fetched= after an interval where the front-end delivered no uops for a period of 64= cycles which was not interrupted by a back-end stall. Available PDIST coun= ters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -343,7 +343,7 @@ "EventName": "FRONTEND_RETIRED.LATENCY_GE_8", "MSRIndex": "0x3F7", "MSRValue": "0x600806", - "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 8 cycles. During thi= s period the front-end delivered no uops. Available PDIST counters: 0", + "PublicDescription": "Counts retired instructions that are deliver= ed to the back-end after a front-end stall of at least 8 cycles. During thi= s period the front-end delivered no uops. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -355,7 +355,7 @@ "EventName": "FRONTEND_RETIRED.MISP_ANT", "MSRIndex": "0x3F7", "MSRValue": "0x9", - "PublicDescription": "ANT retired branches that got just mispredic= ted Available PDIST counters: 0", + "PublicDescription": "ANT retired branches that got just mispredic= ted Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x2", "Unit": "cpu_core" @@ -367,7 +367,7 @@ "EventName": "FRONTEND_RETIRED.MS_FLOWS", "MSRIndex": "0x3F7", "MSRValue": "0x8", - "PublicDescription": "Counts flows delivered by the Microcode Sequ= encer Available PDIST counters: 0", + "PublicDescription": "Counts flows delivered by the Microcode Sequ= encer Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -397,7 +397,7 @@ "EventName": "FRONTEND_RETIRED.STLB_MISS", "MSRIndex": "0x3F7", "MSRValue": "0x15", - "PublicDescription": "Counts retired Instructions that experienced= STLB (2nd level TLB) true miss. Available PDIST counters: 0", + "PublicDescription": "Counts retired Instructions that experienced= STLB (2nd level TLB) true miss. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" @@ -409,7 +409,7 @@ "EventName": "FRONTEND_RETIRED.UNKNOWN_BRANCH", "MSRIndex": "0x3F7", "MSRValue": "0x17", - "PublicDescription": "Number retired branch instructions that caus= ed the front-end to be resteered when it finds the instruction in a fetch l= ine. This is called Unknown Branch which occurs for the first time a branch= instruction is fetched or when the branch is not tracked by the BPU (Branc= h Prediction Unit) anymore. Available PDIST counters: 0", + "PublicDescription": "Number retired branch instructions that caus= ed the front-end to be resteered when it finds the instruction in a fetch l= ine. This is called Unknown Branch which occurs for the first time a branch= instruction is fetched or when the branch is not tracked by the BPU (Branc= h Prediction Unit) anymore. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x3", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/lunarlake/lnl-metrics.json b/to= ols/perf/pmu-events/arch/x86/lunarlake/lnl-metrics.json index 3c740962e63e..06390a72110d 100644 --- a/tools/perf/pmu-events/arch/x86/lunarlake/lnl-metrics.json +++ b/tools/perf/pmu-events/arch/x86/lunarlake/lnl-metrics.json @@ -1,74 +1,46 @@ [ { "BriefDescription": "C10 residency percent per package", - "MetricExpr": "cstate_pkg@c10\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c10\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C10_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, - { - "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", - "MetricGroup": "Power", - "MetricName": "C3_Pkg_Residency", - "ScaleUnit": "100%" - }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, - { - "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", - "MetricGroup": "Power", - "MetricName": "C7_Pkg_Residency", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "C8 residency percent per package", - "MetricExpr": "cstate_pkg@c8\\-residency@ / TSC", - "MetricGroup": "Power", - "MetricName": "C8_Pkg_Residency", - "ScaleUnit": "100%" - }, - { - "BriefDescription": "C9 residency percent per package", - "MetricExpr": "cstate_pkg@c9\\-residency@ / TSC", - "MetricGroup": "Power", - "MetricName": "C9_Pkg_Residency", - "ScaleUnit": "100%" - }, { "BriefDescription": "Percentage of cycles spent in System Manageme= nt Interrupts.", "MetricExpr": "((msr@aperf@ - cycles) / msr@aperf@ if msr@smi@ > 0= else 0)", @@ -555,7 +527,7 @@ }, { "BriefDescription": "Average CPU Utilization", - "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.REF_TSC@ / TSC", + "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.REF_TSC@ / msr@tsc\\,cpu= =3Dcpu_atom@", "MetricName": "tma_info_system_cpu_utilization", "Unit": "cpu_atom" }, @@ -724,6 +696,13 @@ "ScaleUnit": "100%", "Unit": "cpu_atom" }, + { + "BriefDescription": "Uncore frequency per die [GHZ]", + "MetricExpr": "tma_info_system_socket_clks / #num_dies / duration_= time / 1e9", + "MetricGroup": "SoC", + "MetricName": "UNCORE_FREQ", + "Unit": "cpu_core" + }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", "MetricExpr": "cpu_core@UOPS_DISPATCHED.ALU@ / (6 * tma_info_threa= d_clks)", @@ -755,7 +734,7 @@ { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-be\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-be\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BvOB;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", @@ -767,7 +746,7 @@ { "BriefDescription": "This category represents fraction of slots wa= sted due to incorrect speculations", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-bad\\-spec@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-bad\\-spec@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_bad_speculation", "MetricThreshold": "tma_bad_speculation > 0.15", @@ -793,36 +772,36 @@ "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)", "Unit": "cpu_core" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: ", + "Unit": "cpu_core" + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_capacity + tma_l1_latency_dependency + tma_lock_latency + tma_split_= loads + tma_store_fwd_blk)))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_capacity + tma_l1_latency_dependency + tma_lock_latency + tma_split_= loads + tma_store_early_blk + tma_store_fwd_blk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full", "Unit": "cpu_core" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_capacity= + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_sto= re_fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_= bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_latency_c= apacity / (tma_dtlb_load + tma_fb_full + tma_l1_latency_capacity + tma_l1_l= atency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)= ) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma= _l2_bound + tma_l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_dtl= b_load + tma_fb_full + tma_l1_latency_capacity + tma_l1_latency_dependency = + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bou= nd * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_split_loads / (tma_dtlb_load + tma_fb_ful= l + tma_l1_latency_capacity + tma_l1_latency_dependency + tma_lock_latency = + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bou= nd / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_sto= re_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tma_= split_stores + tma_store_latency + tma_streaming_stores)) + tma_memory_boun= d * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store + tma_f= alse_sharing + tma_split_stores + tma_store_latency + tma_streaming_stores)= ))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_capacity= + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_sto= re_early_blk + tma_store_fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tm= a_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound= )) * (tma_l1_latency_capacity / (tma_dtlb_load + tma_fb_full + tma_l1_laten= cy_capacity + tma_l1_latency_dependency + tma_lock_latency + tma_split_load= s + tma_store_early_blk + tma_store_fwd_blk)) + tma_memory_bound * (tma_l1_= bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_= store_bound)) * (tma_lock_latency / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_capacity + tma_l1_latency_dependency + tma_lock_latency + tma_split_= loads + tma_store_early_blk + tma_store_fwd_blk)) + tma_memory_bound * (tma= _l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + = tma_store_bound)) * (tma_split_loads / (tma_dtlb_load + tma_fb_full + tma_l= 1_latency_capacity + tma_l1_latency_dependency + tma_lock_latency + tma_spl= it_loads + tma_store_early_blk + tma_store_fwd_blk)) + tma_memory_bound * (= tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bo= und + tma_store_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_s= haring + tma_split_stores + tma_store_latency + tma_streaming_stores)) + tm= a_memory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2= _bound + tma_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_= store + tma_false_sharing + tma_split_stores + tma_store_latency + tma_stre= aming_stores)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency", "Unit": "cpu_core" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: ", - "Unit": "cpu_core" - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - c= pu_core@INST_RETIRED.REP_ITERATION@ / cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D= 1@) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cl= ears_resteers + tma_mispredicts_resteers * tma_other_mispredicts / tma_bran= ch_mispredicts) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unk= nown_branches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_miss= es + tma_itlb_misses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * t= ma_ms / (tma_dsb + tma_lsd + tma_mite + tma_ms))) - tma_bottleneck_big_code= ", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - c= pu_core@INST_RETIRED.REP_ITERATION@ / cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D= 1@) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cl= ears_resteers + tma_mispredicts_resteers * tma_other_mispredicts / tma_bran= ch_mispredicts) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unk= nown_branches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_miss= es + tma_itlb_misses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottlene= ck_big_code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20", @@ -830,7 +809,7 @@ }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * ((1 - cpu_core@INST_RETIRED.REP_ITERATION@ / = cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_restee= rs * tma_other_mispredicts / tma_branch_mispredicts) / (tma_clears_resteers= + tma_mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers= + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_m= s_switches) + tma_fetch_bandwidth * tma_ms / (tma_dsb + tma_lsd + tma_mite = + tma_ms)) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_bra= nch_mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_n= ukes / tma_other_nukes + tma_core_bound * (tma_serializing_operation + cpu_= core@RS.EMPTY_RESOURCE@ / tma_info_thread_clks * tma_ports_utilized_0) / (t= ma_divider + tma_ports_utilization + tma_serializing_operation) + tma_micro= code_sequencer / (tma_microcode_sequencer + tma_few_uops_instructions) * (t= ma_assists / tma_microcode_sequencer) * tma_heavy_operations)", + "MetricExpr": "100 * ((1 - cpu_core@INST_RETIRED.REP_ITERATION@ / = cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_restee= rs * tma_other_mispredicts / tma_branch_mispredicts) / (tma_clears_resteers= + tma_mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers= + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_m= s_switches) + tma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredict= s / tma_branch_mispredicts * tma_branch_mispredicts + tma_machine_clears * = tma_other_nukes / tma_other_nukes + tma_core_bound * (tma_serializing_opera= tion + cpu_core@RS.EMPTY_RESOURCE@ / tma_info_thread_clks * tma_ports_utili= zed_0) / (tma_divider + tma_ports_utilization + tma_serializing_operation) = + tma_microcode_sequencer / (tma_microcode_sequencer + tma_few_uops_instruc= tions) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -839,7 +818,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", - "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / (tma_dram= _bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (= tma_dtlb_load / (tma_dtlb_load + tma_fb_full + tma_l1_latency_capacity + tm= a_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fw= d_blk)) + tma_memory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bo= und + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_dtlb_store / (= tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_store_latency += tma_streaming_stores)))", + "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / (tma_dram= _bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (= tma_dtlb_load / (tma_dtlb_load + tma_fb_full + tma_l1_latency_capacity + tm= a_l1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_ea= rly_blk + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_= dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound))= * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_split_stores= + tma_store_latency + tma_streaming_stores)))", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", "MetricThreshold": "tma_bottleneck_memory_data_tlbs > 20", @@ -866,7 +845,7 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -883,7 +862,7 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU has wasted due to Branch Misprediction", - "MetricExpr": "cpu_core@topdown\\-br\\-mispredict@ / (cpu_core@top= down\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-re= tiring@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-br\\-mispredict@ / (cpu_core@top= down\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-re= tiring@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BadSpec;BrMispredicts;BvMP;TmaL2;TopdownL2;tma_L2_= group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_specula= tion > 0.15", @@ -1023,7 +1002,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to contested acces= ses", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS@ * min(= cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS@R, 24 * tma_info_system_core_fre= quency) + cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_HITM@ * min(cpu_core@MEM_LO= AD_L3_HIT_RETIRED.XSNP_HITM@R, 25 * tma_info_system_core_frequency)) * (1 += cpu_core@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2= ) / tma_info_thread_clks", "MetricGroup": "BvMS;DataSharing;LockCont;Offcore;Snoop;TopdownL4;= tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", @@ -1076,7 +1054,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to DSB (decoded uop cache) fetch pipe= line", - "MetricExpr": "(cpu_core@IDQ.DSB_UOPS\\,cmask\\=3D0x8\\,inv\\=3D0x= 1@ + cpu_core@IDQ.DSB_UOPS@ / (cpu_core@IDQ.DSB_UOPS@ + cpu_core@IDQ.MITE_U= OPS@) * (cpu_core@IDQ_BUBBLES.CYCLES_0_UOPS_DELIV.CORE@ - cpu_core@IDQ_BUBB= LES.FETCH_LATENCY@)) / tma_info_thread_clks", + "MetricExpr": "(cpu_core@IDQ.DSB_UOPS\\,cmask\\=3D0x8\\,inv\\=3D0x= 1@ / 2 + cpu_core@IDQ.DSB_UOPS@ / (cpu_core@IDQ.DSB_UOPS@ + cpu_core@IDQ.MI= TE_UOPS@) * (cpu_core@IDQ_BUBBLES.STARVATION_CYCLES@ - cpu_core@IDQ_BUBBLES= .FETCH_LATENCY@)) / tma_info_thread_clks", "MetricGroup": "DSB;FetchBW;TopdownL3;tma_L3_group;tma_fetch_bandw= idth_group", "MetricName": "tma_dsb", "MetricThreshold": "tma_dsb > 0.15 & tma_fetch_bandwidth > 0.2", @@ -1130,7 +1108,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -1147,7 +1125,7 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU was stalled due to Frontend latency issues", - "MetricExpr": "cpu_core@topdown\\-fetch\\-lat@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-fetch\\-lat@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Frontend;TmaL2;TopdownL2;tma_L2_group;tma_frontend= _bound_group", "MetricName": "tma_fetch_latency", "MetricThreshold": "tma_fetch_latency > 0.1 & tma_frontend_bound >= 0.15", @@ -1197,7 +1175,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-= point (FP) scalar uops fraction the CPU has retired", - "MetricExpr": "cpu_core@FP_ARITH_INST_RETIRED.SCALAR@ / (tma_retir= ing * tma_info_thread_slots)", + "MetricExpr": "cpu_core@FP_ARITH_OPS_RETIRED.SCALAR@ / (tma_retiri= ng * tma_info_thread_slots)", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_= group;tma_issue2P", "MetricName": "tma_fp_scalar", "MetricThreshold": "tma_fp_scalar > 0.1 & (tma_fp_arith > 0.2 & tm= a_light_operations > 0.6)", @@ -1207,7 +1185,7 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-= point (FP) vector uops fraction the CPU has retired aggregated across all v= ector widths", - "MetricExpr": "cpu_core@FP_ARITH_INST_RETIRED.VECTOR@ / (tma_retir= ing * tma_info_thread_slots)", + "MetricExpr": "cpu_core@FP_ARITH_OPS_RETIRED.VECTOR@ / (tma_retiri= ng * tma_info_thread_slots)", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_= group;tma_issue2P", "MetricName": "tma_fp_vector", "MetricThreshold": "tma_fp_vector > 0.1 & (tma_fp_arith > 0.2 & tm= a_light_operations > 0.6)", @@ -1217,7 +1195,7 @@ }, { "BriefDescription": "This metric approximates arithmetic FP vector= uops fraction the CPU has retired for 128-bit wide vectors", - "MetricExpr": "(cpu_core@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE@= + cpu_core@FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE@) / (tma_retiring * tm= a_info_thread_slots)", + "MetricExpr": "(cpu_core@FP_ARITH_OPS_RETIRED.128B_PACKED_DOUBLE@ = + cpu_core@FP_ARITH_OPS_RETIRED.128B_PACKED_SINGLE@) / (tma_retiring * tma_= info_thread_slots)", "MetricGroup": "Compute;Flops;TopdownL5;tma_L5_group;tma_fp_vector= _group;tma_issue2P", "MetricName": "tma_fp_vector_128b", "MetricThreshold": "tma_fp_vector_128b > 0.1 & (tma_fp_vector > 0.= 1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6))", @@ -1227,7 +1205,7 @@ }, { "BriefDescription": "This metric approximates arithmetic FP vector= uops fraction the CPU has retired for 256-bit wide vectors", - "MetricExpr": "cpu_core@FP_ARITH_INST_RETIRED.VECTOR\\,umask\\=3D0= x30@ / (tma_retiring * tma_info_thread_slots)", + "MetricExpr": "cpu_core@FP_ARITH_OPS_RETIRED.VECTOR\\,umask\\=3D0x= 30@ / (tma_retiring * tma_info_thread_slots)", "MetricGroup": "Compute;Flops;TopdownL5;tma_L5_group;tma_fp_vector= _group;tma_issue2P", "MetricName": "tma_fp_vector_256b", "MetricThreshold": "tma_fp_vector_256b > 0.1 & (tma_fp_vector > 0.= 1 & (tma_fp_arith > 0.2 & tma_light_operations > 0.6))", @@ -1238,7 +1216,7 @@ { "BriefDescription": "This category represents fraction of slots wh= ere the processor's Frontend undersupplies its Backend", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-fe\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-fe\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BvFB;BvIO;Default;PGO;TmaL1;TopdownL1;tma_L1_group= ", "MetricName": "tma_frontend_bound", "MetricThreshold": "tma_frontend_bound > 0.15", @@ -1259,7 +1237,7 @@ }, { "BriefDescription": "This metric represents fraction of slots wher= e the CPU was retiring heavy-weight operations -- instructions that require= two or more uops or micro-coded sequences", - "MetricExpr": "cpu_core@topdown\\-heavy\\-ops@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-heavy\\-ops@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Retire;TmaL2;TopdownL2;tma_L2_group;tma_retiring_g= roup", "MetricName": "tma_heavy_operations", "MetricThreshold": "tma_heavy_operations > 0.1", @@ -1437,7 +1415,7 @@ }, { "BriefDescription": "Floating Point Operations Per Cycle", - "MetricExpr": "(cpu_core@FP_ARITH_INST_RETIRED.SCALAR@ + 2 * cpu_c= ore@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE@ + 4 * cpu_core@FP_ARITH_INST_= RETIRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE@) = / tma_info_thread_clks", + "MetricExpr": "(cpu_core@FP_ARITH_OPS_RETIRED.SCALAR@ + 2 * cpu_co= re@FP_ARITH_OPS_RETIRED.128B_PACKED_DOUBLE@ + 4 * cpu_core@FP_ARITH_OPS_RET= IRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_OPS_RETIRED.256B_PACKED_SINGLE@) / tm= a_info_thread_clks", "MetricGroup": "Flops;Ret", "MetricName": "tma_info_core_flopc", "Unit": "cpu_core" @@ -1578,7 +1556,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic instruction (l= ower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_INS= T_RETIRED.SCALAR@ + cpu_core@FP_ARITH_INST_RETIRED.VECTOR@)", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_OPS= _RETIRED.SCALAR@ + cpu_core@FP_ARITH_OPS_RETIRED.VECTOR@)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_iparith", "MetricThreshold": "tma_info_inst_mix_iparith < 10", @@ -1587,7 +1565,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic AVX/SSE 128-bi= t instruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_INS= T_RETIRED.128B_PACKED_DOUBLE@ + cpu_core@FP_ARITH_INST_RETIRED.128B_PACKED_= SINGLE@)", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_OPS= _RETIRED.128B_PACKED_DOUBLE@ + cpu_core@FP_ARITH_OPS_RETIRED.128B_PACKED_SI= NGLE@)", "MetricGroup": "Flops;FpVector;InsType", "MetricName": "tma_info_inst_mix_iparith_avx128", "MetricThreshold": "tma_info_inst_mix_iparith_avx128 < 10", @@ -1596,7 +1574,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic AVX* 256-bit i= nstruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_INS= T_RETIRED.256B_PACKED_DOUBLE@ + cpu_core@FP_ARITH_INST_RETIRED.256B_PACKED_= SINGLE@)", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_OPS= _RETIRED.256B_PACKED_DOUBLE@ + cpu_core@FP_ARITH_OPS_RETIRED.256B_PACKED_SI= NGLE@)", "MetricGroup": "Flops;FpVector;InsType", "MetricName": "tma_info_inst_mix_iparith_avx256", "MetricThreshold": "tma_info_inst_mix_iparith_avx256 < 10", @@ -1605,7 +1583,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic Scalar Double-= Precision instruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FP_ARITH_INST= _RETIRED.SCALAR_DOUBLE@", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FP_ARITH_OPS_= RETIRED.SCALAR_DOUBLE@", "MetricGroup": "Flops;FpScalar;InsType", "MetricName": "tma_info_inst_mix_iparith_scalar_dp", "MetricThreshold": "tma_info_inst_mix_iparith_scalar_dp < 10", @@ -1614,7 +1592,7 @@ }, { "BriefDescription": "Instructions per FP Arithmetic Scalar Single-= Precision instruction (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FP_ARITH_INST= _RETIRED.SCALAR_SINGLE@", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@FP_ARITH_OPS_= RETIRED.SCALAR_SINGLE@", "MetricGroup": "Flops;FpScalar;InsType", "MetricName": "tma_info_inst_mix_iparith_scalar_sp", "MetricThreshold": "tma_info_inst_mix_iparith_scalar_sp < 10", @@ -1639,7 +1617,7 @@ }, { "BriefDescription": "Instructions per Floating Point (FP) Operatio= n (lower number means higher occurrence rate)", - "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_INS= T_RETIRED.SCALAR@ + 2 * cpu_core@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE@ = + 4 * cpu_core@FP_ARITH_INST_RETIRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_INST_= RETIRED.256B_PACKED_SINGLE@)", + "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / (cpu_core@FP_ARITH_OPS= _RETIRED.SCALAR@ + 2 * cpu_core@FP_ARITH_OPS_RETIRED.128B_PACKED_DOUBLE@ + = 4 * cpu_core@FP_ARITH_OPS_RETIRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_OPS_RETI= RED.256B_PACKED_SINGLE@)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_ipflop", "MetricThreshold": "tma_info_inst_mix_ipflop < 10", @@ -1694,7 +1672,7 @@ }, { "BriefDescription": "Average per-thread data fill bandwidth to the= L1 data cache [GB / sec]", - "MetricExpr": "64 * cpu_core@L1D.REPLACEMENT@ / 1e9 / tma_info_sys= tem_time", + "MetricExpr": "64 * cpu_core@L1D.L1_REPLACEMENT@ / 1e9 / tma_info_= system_time", "MetricGroup": "Mem;MemoryBW", "MetricName": "tma_info_memory_l1d_cache_fill_bw", "Unit": "cpu_core" @@ -1706,6 +1684,13 @@ "MetricName": "tma_info_memory_l1dl0_cache_fill_bw", "Unit": "cpu_core" }, + { + "BriefDescription": "L0 cache true misses per kilo instruction for= retired demand loads", + "MetricExpr": "1e3 * (cpu_core@MEM_LOAD_RETIRED.L1_MISS@ + cpu_cor= e@MEM_LOAD_RETIRED.L1_HIT_L1@) / cpu_core@INST_RETIRED.ANY@", + "MetricGroup": "CacheHits;Mem", + "MetricName": "tma_info_memory_l1dl0_mpki", + "Unit": "cpu_core" + }, { "BriefDescription": "L1 cache true misses per kilo instruction for= retired demand loads", "MetricExpr": "1e3 * cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / cpu_core= @INST_RETIRED.ANY@", @@ -1921,6 +1906,13 @@ "MetricName": "tma_info_pipeline_fetch_mite", "Unit": "cpu_core" }, + { + "BriefDescription": "Average number of uops fetched from MS per cy= cle", + "MetricExpr": "cpu_core@IDQ.MS_UOPS@ / cpu_core@IDQ.MS_UOPS\\,cmas= k\\=3D1@", + "MetricGroup": "Fed;FetchLat;MicroSeq", + "MetricName": "tma_info_pipeline_fetch_ms", + "Unit": "cpu_core" + }, { "BriefDescription": "Instructions per a microcode Assist invocatio= n", "MetricExpr": "cpu_core@INST_RETIRED.ANY@ / cpu_core@ASSISTS.ANY@", @@ -1955,7 +1947,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc\\,cpu= =3Dcpu_core@ / 1e9 / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency", "Unit": "cpu_core" @@ -1969,14 +1961,22 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.REF_TSC@ / TSC", + "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.REF_TSC@ / msr@tsc\\,cpu= =3Dcpu_core@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized", "Unit": "cpu_core" }, + { + "BriefDescription": "Average external Memory Bandwidth Use for rea= ds and writes [GB / sec]", + "MetricExpr": "32 * UNC_M_TOTAL_DATA / 1e9 / tma_info_system_time", + "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", + "MetricName": "tma_info_system_dram_bw_use", + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full", + "Unit": "cpu_core" + }, { "BriefDescription": "Giga Floating Point Operations Per Second", - "MetricExpr": "(cpu_core@FP_ARITH_INST_RETIRED.SCALAR@ + 2 * cpu_c= ore@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE@ + 4 * cpu_core@FP_ARITH_INST_= RETIRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE@) = / 1e9 / tma_info_system_time", + "MetricExpr": "(cpu_core@FP_ARITH_OPS_RETIRED.SCALAR@ + 2 * cpu_co= re@FP_ARITH_OPS_RETIRED.128B_PACKED_DOUBLE@ + 4 * cpu_core@FP_ARITH_OPS_RET= IRED.4_FLOPS@ + 8 * cpu_core@FP_ARITH_OPS_RETIRED.256B_PACKED_SINGLE@) / 1e= 9 / tma_info_system_time", "MetricGroup": "Cor;Flops;HPC", "MetricName": "tma_info_system_gflops", "PublicDescription": "Giga Floating Point Operations Per Second. A= ggregate across all supported options of: FP precisions, scalar and vector = instructions, vector-width", @@ -2020,6 +2020,13 @@ "MetricName": "tma_info_system_power", "Unit": "cpu_core" }, + { + "BriefDescription": "Socket actual clocks when any core is active = on that socket", + "MetricExpr": "UNC_CLOCK.SOCKET", + "MetricGroup": "SoC", + "MetricName": "tma_info_system_socket_clks", + "Unit": "cpu_core" + }, { "BriefDescription": "Run duration time in seconds", "MetricExpr": "duration_time", @@ -2035,6 +2042,13 @@ "MetricName": "tma_info_system_turbo_utilization", "Unit": "cpu_core" }, + { + "BriefDescription": "Measured Average Uncore Frequency for the SoC= [GHz]", + "MetricExpr": "tma_info_system_socket_clks / 1e9 / tma_info_system= _time", + "MetricGroup": "SoC", + "MetricName": "tma_info_system_uncore_frequency", + "Unit": "cpu_core" + }, { "BriefDescription": "Per-Logical Processor actual clocks when the = Logical Processor is active.", "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.THREAD@", @@ -2156,12 +2170,12 @@ "Unit": "cpu_core" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", - "MetricExpr": "4 * cpu_core@DEPENDENT_LOADS.ANY@ / tma_info_thread= _clks", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", + "MetricExpr": "4 * cpu_core@DEPENDENT_LOADS.ANY\\,cmask\\=3D1@ / t= ma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_UOPS_RETIRED.L1_HIT_PS", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_UOPS_RETIRED.L1_HIT_PS", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2177,7 +2191,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles wit= h demand load accesses that hit the L2 cache under unloaded scenarios (poss= ibly L2 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu_core@MEM_LOAD_RETIRED.L2_HIT@ * min(cpu_core@ME= M_LOAD_RETIRED.L2_HIT@R, 3 * tma_info_system_core_frequency) * (1 + cpu_cor= e@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2) / tma_= info_thread_clks", "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_l2_bound_grou= p", "MetricName": "tma_l2_hit_latency", @@ -2198,12 +2211,11 @@ }, { "BriefDescription": "This metric estimates fraction of cycles with= demand load accesses that hit the L3 cache under unloaded scenarios (possi= bly L3 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu_core@MEM_LOAD_RETIRED.L3_HIT@ * min(cpu_core@ME= M_LOAD_RETIRED.L3_HIT@R, 9 * tma_info_system_core_frequency) * (1 + cpu_cor= e@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2) / tma_= info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2285,6 +2297,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU spent handling cache misses due to lock operations", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu_core@MEM_INST_RETIRED.LOCK_LOADS@ * cpu_core@ME= M_INST_RETIRED.LOCK_LOADS@R / tma_info_thread_clks", "MetricGroup": "LockCont;Offcore;TopdownL4;tma_L4_group;tma_issueR= FO;tma_l1_bound_group", "MetricName": "tma_lock_latency", @@ -2295,7 +2308,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to LSD (Loop Stream Detector) unit", - "MetricExpr": "cpu_core@LSD.UOPS\\,cmask\\=3D0x8\\,inv\\=3D0x1@ / = tma_info_thread_clks", + "MetricExpr": "cpu_core@LSD.UOPS\\,cmask\\=3D0x8\\,inv\\=3D0x1@ / = tma_info_thread_clks / 2", "MetricGroup": "FetchBW;LSD;TopdownL3;tma_L3_group;tma_fetch_bandw= idth_group", "MetricName": "tma_lsd", "MetricThreshold": "tma_lsd > 0.15 & tma_fetch_bandwidth > 0.2", @@ -2320,7 +2333,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2330,13 +2343,13 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%", "Unit": "cpu_core" }, { "BriefDescription": "This metric represents fraction of slots the = Memory subsystem within the Backend was a bottleneck", - "MetricExpr": "cpu_core@topdown\\-mem\\-bound@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-mem\\-bound@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Backend;TmaL2;TopdownL2;tma_L2_group;tma_backend_b= ound_group", "MetricName": "tma_memory_bound", "MetricThreshold": "tma_memory_bound > 0.2 & tma_backend_bound > 0= .2", @@ -2347,7 +2360,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to LFENCE Instructions.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * cpu_core@MISC2_RETIRED.LFENCE@ / tma_info_thre= ad_clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_memory_fence", @@ -2386,7 +2398,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the MITE pipeline (the legacy deco= de pipeline)", - "MetricExpr": "(cpu_core@IDQ.MITE_UOPS\\,cmask\\=3D0x8\\,inv\\=3D0= x1@ / 2 + cpu_core@IDQ.MITE_UOPS@ / (cpu_core@IDQ.DSB_UOPS@ + cpu_core@IDQ.= MITE_UOPS@) * (cpu_core@IDQ_BUBBLES.CYCLES_0_UOPS_DELIV.CORE@ - cpu_core@ID= Q_BUBBLES.FETCH_LATENCY@)) / tma_info_thread_clks", + "MetricExpr": "(cpu_core@IDQ.MITE_UOPS\\,cmask\\=3D0x8\\,inv\\=3D0= x1@ / 2 + cpu_core@IDQ.MITE_UOPS@ / (cpu_core@IDQ.DSB_UOPS@ + cpu_core@IDQ.= MITE_UOPS@) * (cpu_core@IDQ_BUBBLES.STARVATION_CYCLES@ - cpu_core@IDQ_BUBBL= ES.FETCH_LATENCY@)) / tma_info_thread_clks", "MetricGroup": "DSBmiss;FetchBW;TopdownL3;tma_L3_group;tma_fetch_b= andwidth_group", "MetricName": "tma_mite", "MetricThreshold": "tma_mite > 0.1 & tma_fetch_bandwidth > 0.2", @@ -2406,7 +2418,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "cpu_core@IDQ.MS_CYCLES_ANY@ / tma_info_thread_clks", + "MetricExpr": "cpu_core@IDQ.MS_CYCLES_ANY@ / tma_info_thread_clks = / 1.8", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -2445,7 +2457,8 @@ }, { "BriefDescription": "This metric represents the remaining light uo= ps fraction the CPU has executed - remaining means not covered by other sib= ling nodes", - "MetricExpr": "max(0, tma_light_operations - (tma_x87_use + (cpu_c= ore@FP_ARITH_INST_RETIRED.SCALAR@ + cpu_core@FP_ARITH_INST_RETIRED.VECTOR@)= / (tma_retiring * tma_info_thread_slots) + (cpu_core@INT_VEC_RETIRED.ADD_1= 28@ + cpu_core@INT_VEC_RETIRED.VNNI_128@ + cpu_core@INT_VEC_RETIRED.ADD_256= @ + cpu_core@INT_VEC_RETIRED.MUL_256@ + cpu_core@INT_VEC_RETIRED.VNNI_256@)= / (tma_retiring * tma_info_thread_slots) + tma_memory_operations + tma_fus= ed_instructions + tma_non_fused_branches))", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "max(0, tma_light_operations - (tma_x87_use + (cpu_c= ore@FP_ARITH_OPS_RETIRED.SCALAR@ + cpu_core@FP_ARITH_OPS_RETIRED.VECTOR@) /= (tma_retiring * tma_info_thread_slots) + (cpu_core@INT_VEC_RETIRED.ADD_128= @ + cpu_core@INT_VEC_RETIRED.VNNI_128@ + cpu_core@INT_VEC_RETIRED.ADD_256@ = + cpu_core@INT_VEC_RETIRED.MUL_256@ + cpu_core@INT_VEC_RETIRED.VNNI_256@) /= (tma_retiring * tma_info_thread_slots) + tma_memory_operations + tma_fused= _instructions + tma_non_fused_branches))", "MetricGroup": "Pipeline;TopdownL3;tma_L3_group;tma_light_operatio= ns_group", "MetricName": "tma_other_light_ops", "MetricThreshold": "tma_other_light_ops > 0.3 & tma_light_operatio= ns > 0.6", @@ -2483,6 +2496,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "((cpu_core@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ + (cpu_c= ore@EXE_ACTIVITY.1_PORTS_UTIL@ + tma_retiring * cpu_core@EXE_ACTIVITY.2_3_P= ORTS_UTIL@)) / tma_info_thread_clks if cpu_core@ARITH.DIV_ACTIVE@ < cpu_cor= e@CYCLE_ACTIVITY.STALLS_TOTAL@ - cpu_core@EXE_ACTIVITY.BOUND_ON_LOADS@ else= (cpu_core@EXE_ACTIVITY.1_PORTS_UTIL@ + tma_retiring * cpu_core@EXE_ACTIVIT= Y.2_3_PORTS_UTIL@) / tma_info_thread_clks)", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", @@ -2493,6 +2507,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "cpu_core@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ / tma_info= _thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", @@ -2503,6 +2518,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "cpu_core@EXE_ACTIVITY.1_PORTS_UTIL@ / tma_info_thre= ad_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", @@ -2513,7 +2529,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@EXE_ACTIVITY.2_PORTS_UTIL@ / tma_info_thre= ad_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", @@ -2524,7 +2539,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 3 or more uops per cycle on all execution ports (Logical= Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@UOPS_EXECUTED.CYCLES_GE_3@ / tma_info_thre= ad_clks", "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_ut= ilization_group", "MetricName": "tma_ports_utilized_3m", @@ -2545,7 +2559,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-retiring@ / (cpu_core@topdown\\-= fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@= + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-retiring@ / (cpu_core@topdown\\-= fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@= + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -2576,7 +2590,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to PAUSE Instructions", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.PAUSE@ / tma_info_thread_= clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_slow_pause", @@ -2611,7 +2624,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2625,6 +2638,15 @@ "ScaleUnit": "100%", "Unit": "cpu_core" }, + { + "BriefDescription": "This metric estimates clocks wasted due to lo= ads blocked due to unknown store address (did not do memory disambiguation)= or due to unknown store data", + "MetricExpr": "7 * cpu_core@LD_BLOCKS.STORE_EARLY\\,cmask\\=3D1@ /= tma_info_thread_clks", + "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", + "MetricName": "tma_store_early_blk", + "MetricThreshold": "tma_store_early_blk > 0.2", + "ScaleUnit": "100%", + "Unit": "cpu_core" + }, { "BriefDescription": "This metric roughly estimates fraction of cyc= les when the memory subsystem had loads blocked since they could not forwar= d data from earlier (in program order) overlapping stores", "MetricExpr": "13 * cpu_core@LD_BLOCKS.STORE_FORWARD@ / tma_info_t= hread_clks", diff --git a/tools/perf/pmu-events/arch/x86/lunarlake/memory.json b/tools/p= erf/pmu-events/arch/x86/lunarlake/memory.json index 8021a1c7dd3b..25021cb76f61 100644 --- a/tools/perf/pmu-events/arch/x86/lunarlake/memory.json +++ b/tools/perf/pmu-events/arch/x86/lunarlake/memory.json @@ -163,7 +163,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_1024", "MSRIndex": "0x3F6", "MSRValue": "0x400", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 1024 cycles. Reporte= d latency may be longer than just the memory latency. Available PDIST count= ers: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 1024 cycles. Reporte= d latency may be longer than just the memory latency.", "SampleAfterValue": "53", "UMask": "0x1", "Unit": "cpu_core" @@ -176,7 +176,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_128", "MSRIndex": "0x3F6", "MSRValue": "0x80", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 128 cycles. Reported= latency may be longer than just the memory latency. Available PDIST counte= rs: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 128 cycles. Reported= latency may be longer than just the memory latency.", "SampleAfterValue": "1009", "UMask": "0x1", "Unit": "cpu_core" @@ -189,7 +189,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_16", "MSRIndex": "0x3F6", "MSRValue": "0x10", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 16 cycles. Reported = latency may be longer than just the memory latency. Available PDIST counter= s: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 16 cycles. Reported = latency may be longer than just the memory latency.", "SampleAfterValue": "20011", "UMask": "0x1", "Unit": "cpu_core" @@ -202,7 +202,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_2048", "MSRIndex": "0x3F6", "MSRValue": "0x800", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 2048 cycles. Reporte= d latency may be longer than just the memory latency. Available PDIST count= ers: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 2048 cycles. Reporte= d latency may be longer than just the memory latency.", "SampleAfterValue": "23", "UMask": "0x1", "Unit": "cpu_core" @@ -215,7 +215,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_256", "MSRIndex": "0x3F6", "MSRValue": "0x100", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 256 cycles. Reported= latency may be longer than just the memory latency. Available PDIST counte= rs: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 256 cycles. Reported= latency may be longer than just the memory latency.", "SampleAfterValue": "503", "UMask": "0x1", "Unit": "cpu_core" @@ -228,7 +228,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_32", "MSRIndex": "0x3F6", "MSRValue": "0x20", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 32 cycles. Reported = latency may be longer than just the memory latency. Available PDIST counter= s: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 32 cycles. Reported = latency may be longer than just the memory latency.", "SampleAfterValue": "100007", "UMask": "0x1", "Unit": "cpu_core" @@ -241,7 +241,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_4", "MSRIndex": "0x3F6", "MSRValue": "0x4", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 4 cycles. Reported l= atency may be longer than just the memory latency. Available PDIST counters= : 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 4 cycles. Reported l= atency may be longer than just the memory latency.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -254,7 +254,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_512", "MSRIndex": "0x3F6", "MSRValue": "0x200", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 512 cycles. Reported= latency may be longer than just the memory latency. Available PDIST counte= rs: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 512 cycles. Reported= latency may be longer than just the memory latency.", "SampleAfterValue": "101", "UMask": "0x1", "Unit": "cpu_core" @@ -267,7 +267,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_64", "MSRIndex": "0x3F6", "MSRValue": "0x40", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 64 cycles. Reported = latency may be longer than just the memory latency. Available PDIST counter= s: 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 64 cycles. Reported = latency may be longer than just the memory latency.", "SampleAfterValue": "2003", "UMask": "0x1", "Unit": "cpu_core" @@ -280,7 +280,7 @@ "EventName": "MEM_TRANS_RETIRED.LOAD_LATENCY_GT_8", "MSRIndex": "0x3F6", "MSRValue": "0x8", - "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 8 cycles. Reported l= atency may be longer than just the memory latency. Available PDIST counters= : 0", + "PublicDescription": "Counts randomly selected loads when the late= ncy from first dispatch to completion is greater than 8 cycles. Reported l= atency may be longer than just the memory latency.", "SampleAfterValue": "50021", "UMask": "0x1", "Unit": "cpu_core" @@ -291,7 +291,7 @@ "Data_LA": "1", "EventCode": "0xcd", "EventName": "MEM_TRANS_RETIRED.STORE_SAMPLE", - "PublicDescription": "Counts Retired memory accesses with at least= 1 store operation. This PEBS event is the precisely-distributed (PDist) tr= igger covering all stores uops for sampling by the PEBS Store Latency Facil= ity. The facility is described in Intel SDM Volume 3 section 19.9.8 Availab= le PDIST counters: 0", + "PublicDescription": "Counts Retired memory accesses with at least= 1 store operation. This PEBS event is the precisely-distributed (PDist) tr= igger covering all stores uops for sampling by the PEBS Store Latency Facil= ity. The facility is described in Intel SDM Volume 3 section 19.9.8 Availab= le PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json b/tools= /perf/pmu-events/arch/x86/lunarlake/pipeline.json index 6ac410510628..cdaa01e9a57d 100644 --- a/tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json +++ b/tools/perf/pmu-events/arch/x86/lunarlake/pipeline.json @@ -110,7 +110,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.ALL_BRANCHES", - "PublicDescription": "Counts all branch instructions retired. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts all branch instructions retired. Avai= lable PDIST counters: 0,1", "SampleAfterValue": "400009", "Unit": "cpu_core" }, @@ -128,7 +128,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND", - "PublicDescription": "Counts conditional branch instructions retir= ed. Available PDIST counters: 0", + "PublicDescription": "Counts conditional branch instructions retir= ed. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x111", "Unit": "cpu_core" @@ -147,7 +147,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND_NTAKEN", - "PublicDescription": "Counts not taken branch instructions retired= . Available PDIST counters: 0", + "PublicDescription": "Counts not taken branch instructions retired= . Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x10", "Unit": "cpu_core" @@ -166,7 +166,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND_TAKEN", - "PublicDescription": "Counts taken conditional branch instructions= retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken conditional branch instructions= retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x101", "Unit": "cpu_core" @@ -176,7 +176,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND_TAKEN_BWD", - "PublicDescription": "Counts taken backward conditional branch ins= tructions retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken backward conditional branch ins= tructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x1", "Unit": "cpu_core" @@ -186,7 +186,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND_TAKEN_FWD", - "PublicDescription": "Counts taken forward conditional branch inst= ructions retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken forward conditional branch inst= ructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x102", "Unit": "cpu_core" @@ -205,7 +205,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.FAR_BRANCH", - "PublicDescription": "Counts far branch instructions retired. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts far branch instructions retired. Avai= lable PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x40", "Unit": "cpu_core" @@ -224,7 +224,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.INDIRECT", - "PublicDescription": "Counts near indirect branch instructions ret= ired excluding returns. TSX abort is an indirect branch. Available PDIST co= unters: 0", + "PublicDescription": "Counts near indirect branch instructions ret= ired excluding returns. TSX abort is an indirect branch. Available PDIST co= unters: 0,1", "SampleAfterValue": "100003", "UMask": "0x80", "Unit": "cpu_core" @@ -261,7 +261,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_CALL", - "PublicDescription": "Counts both direct and indirect near call in= structions retired. Available PDIST counters: 0", + "PublicDescription": "Counts both direct and indirect near call in= structions retired. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x2", "Unit": "cpu_core" @@ -280,13 +280,13 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_RETURN", - "PublicDescription": "Counts return instructions retired. Availabl= e PDIST counters: 0", + "PublicDescription": "Counts return instructions retired. Availabl= e PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x8", "Unit": "cpu_core" }, { - "BriefDescription": "Counts the number of taken branch instruction= s retired", + "BriefDescription": "Counts the number of near taken branch instru= ctions retired", "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_TAKEN", @@ -299,7 +299,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_TAKEN", - "PublicDescription": "Counts taken branch instructions retired. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts taken branch instructions retired. Av= ailable PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x20", "Unit": "cpu_core" @@ -327,7 +327,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.ALL_BRANCHES", - "PublicDescription": "Counts all the retired branch instructions t= hat were mispredicted by the processor. A branch misprediction occurs when = the processor incorrectly predicts the destination of the branch. When the= misprediction is discovered at execution, all the instructions executed in= the wrong (speculative) path must be discarded, and the processor must sta= rt fetching from the correct path. Available PDIST counters: 0", + "PublicDescription": "Counts all the retired branch instructions t= hat were mispredicted by the processor. A branch misprediction occurs when = the processor incorrectly predicts the destination of the branch. When the= misprediction is discovered at execution, all the instructions executed in= the wrong (speculative) path must be discarded, and the processor must sta= rt fetching from the correct path. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "Unit": "cpu_core" }, @@ -336,7 +336,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.ALL_BRANCHES_COST", - "PublicDescription": "All mispredicted branch instructions retired= . This precise event may be used to get the misprediction cost via the Reti= re_Latency field of PEBS. It fires on the instruction that immediately foll= ows the mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "All mispredicted branch instructions retired= . This precise event may be used to get the misprediction cost via the Reti= re_Latency field of PEBS. It fires on the instruction that immediately foll= ows the mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x44", "Unit": "cpu_core" @@ -355,7 +355,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND", - "PublicDescription": "Counts mispredicted conditional branch instr= uctions retired. Available PDIST counters: 0", + "PublicDescription": "Counts mispredicted conditional branch instr= uctions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x111", "Unit": "cpu_core" @@ -365,7 +365,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_COST", - "PublicDescription": "Mispredicted conditional branch instructions= retired. This precise event may be used to get the misprediction cost via = the Retire_Latency field of PEBS. It fires on the instruction that immediat= ely follows the mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted conditional branch instructions= retired. This precise event may be used to get the misprediction cost via = the Retire_Latency field of PEBS. It fires on the instruction that immediat= ely follows the mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x151", "Unit": "cpu_core" @@ -384,7 +384,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_NTAKEN", - "PublicDescription": "Counts the number of conditional branch inst= ructions retired that were mispredicted and the branch direction was not ta= ken. Available PDIST counters: 0", + "PublicDescription": "Counts the number of conditional branch inst= ructions retired that were mispredicted and the branch direction was not ta= ken. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x10", "Unit": "cpu_core" @@ -394,7 +394,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_NTAKEN_COST", - "PublicDescription": "Mispredicted non-taken conditional branch in= structions retired. This precise event may be used to get the misprediction= cost via the Retire_Latency field of PEBS. It fires on the instruction tha= t immediately follows the mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted non-taken conditional branch in= structions retired. This precise event may be used to get the misprediction= cost via the Retire_Latency field of PEBS. It fires on the instruction tha= t immediately follows the mispredicted branch. Available PDIST counters: 0,= 1", "SampleAfterValue": "400009", "UMask": "0x50", "Unit": "cpu_core" @@ -413,7 +413,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN", - "PublicDescription": "Counts taken conditional mispredicted branch= instructions retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken conditional mispredicted branch= instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x101", "Unit": "cpu_core" @@ -423,7 +423,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN_BWD", - "PublicDescription": "Counts taken backward conditional mispredict= ed branch instructions retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken backward conditional mispredict= ed branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x1", "Unit": "cpu_core" @@ -433,7 +433,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN_BWD_COST", - "PublicDescription": "number of branch instructions retired that w= ere mispredicted and taken backward. This precise event may be used to get = the misprediction cost via the Retire_Latency field of PEBS. It fires on th= e instruction that immediately follows the mispredicted branch. Available P= DIST counters: 0", + "PublicDescription": "number of branch instructions retired that w= ere mispredicted and taken backward. This precise event may be used to get = the misprediction cost via the Retire_Latency field of PEBS. It fires on th= e instruction that immediately follows the mispredicted branch. Available P= DIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x8001", "Unit": "cpu_core" @@ -443,7 +443,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN_COST", - "PublicDescription": "Mispredicted taken conditional branch instru= ctions retired. This precise event may be used to get the misprediction cos= t via the Retire_Latency field of PEBS. It fires on the instruction that im= mediately follows the mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted taken conditional branch instru= ctions retired. This precise event may be used to get the misprediction cos= t via the Retire_Latency field of PEBS. It fires on the instruction that im= mediately follows the mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x141", "Unit": "cpu_core" @@ -453,7 +453,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN_FWD", - "PublicDescription": "Counts taken forward conditional mispredicte= d branch instructions retired. Available PDIST counters: 0", + "PublicDescription": "Counts taken forward conditional mispredicte= d branch instructions retired. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "Unit": "cpu_core" }, @@ -462,7 +462,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.COND_TAKEN_FWD_COST", - "PublicDescription": "number of branch instructions retired that w= ere mispredicted and taken forward. This precise event may be used to get t= he misprediction cost via the Retire_Latency field of PEBS. It fires on the= instruction that immediately follows the mispredicted branch. Available PD= IST counters: 0", + "PublicDescription": "number of branch instructions retired that w= ere mispredicted and taken forward. This precise event may be used to get t= he misprediction cost via the Retire_Latency field of PEBS. It fires on the= instruction that immediately follows the mispredicted branch. Available PD= IST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x8002", "Unit": "cpu_core" @@ -481,7 +481,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.INDIRECT", - "PublicDescription": "Counts miss-predicted near indirect branch i= nstructions retired excluding returns. TSX abort is an indirect branch. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts miss-predicted near indirect branch i= nstructions retired excluding returns. TSX abort is an indirect branch. Ava= ilable PDIST counters: 0,1", "SampleAfterValue": "100003", "UMask": "0x80", "Unit": "cpu_core" @@ -500,7 +500,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.INDIRECT_CALL", - "PublicDescription": "Counts retired mispredicted indirect (near t= aken) CALL instructions, including both register and memory indirect. Avail= able PDIST counters: 0", + "PublicDescription": "Counts retired mispredicted indirect (near t= aken) CALL instructions, including both register and memory indirect. Avail= able PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x2", "Unit": "cpu_core" @@ -510,7 +510,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.INDIRECT_CALL_COST", - "PublicDescription": "Mispredicted indirect CALL retired. This pre= cise event may be used to get the misprediction cost via the Retire_Latency= field of PEBS. It fires on the instruction that immediately follows the mi= spredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted indirect CALL retired. This pre= cise event may be used to get the misprediction cost via the Retire_Latency= field of PEBS. It fires on the instruction that immediately follows the mi= spredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x42", "Unit": "cpu_core" @@ -520,7 +520,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.INDIRECT_COST", - "PublicDescription": "Mispredicted near indirect branch instructio= ns retired (excluding returns). This precise event may be used to get the m= isprediction cost via the Retire_Latency field of PEBS. It fires on the ins= truction that immediately follows the mispredicted branch. Available PDIST = counters: 0", + "PublicDescription": "Mispredicted near indirect branch instructio= ns retired (excluding returns). This precise event may be used to get the m= isprediction cost via the Retire_Latency field of PEBS. It fires on the ins= truction that immediately follows the mispredicted branch. Available PDIST = counters: 0,1", "SampleAfterValue": "100003", "UMask": "0xc0", "Unit": "cpu_core" @@ -548,7 +548,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.NEAR_TAKEN", - "PublicDescription": "Counts number of near branch instructions re= tired that were mispredicted and taken. Available PDIST counters: 0", + "PublicDescription": "Counts number of near branch instructions re= tired that were mispredicted and taken. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x20", "Unit": "cpu_core" @@ -558,7 +558,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.NEAR_TAKEN_COST", - "PublicDescription": "Mispredicted taken near branch instructions = retired. This precise event may be used to get the misprediction cost via t= he Retire_Latency field of PEBS. It fires on the instruction that immediate= ly follows the mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted taken near branch instructions = retired. This precise event may be used to get the misprediction cost via t= he Retire_Latency field of PEBS. It fires on the instruction that immediate= ly follows the mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "400009", "UMask": "0x60", "Unit": "cpu_core" @@ -568,7 +568,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.RET", - "PublicDescription": "This is a non-precise version (that is, does= not use PEBS) of the event that counts mispredicted return instructions re= tired. Available PDIST counters: 0", + "PublicDescription": "This is a non-precise version (that is, does= not use PEBS) of the event that counts mispredicted return instructions re= tired. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x8", "Unit": "cpu_core" @@ -587,7 +587,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc5", "EventName": "BR_MISP_RETIRED.RET_COST", - "PublicDescription": "Mispredicted ret instructions retired. This = precise event may be used to get the misprediction cost via the Retire_Late= ncy field of PEBS. It fires on the instruction that immediately follows the= mispredicted branch. Available PDIST counters: 0", + "PublicDescription": "Mispredicted ret instructions retired. This = precise event may be used to get the misprediction cost via the Retire_Late= ncy field of PEBS. It fires on the instruction that immediately follows the= mispredicted branch. Available PDIST counters: 0,1", "SampleAfterValue": "100007", "UMask": "0x48", "Unit": "cpu_core" @@ -906,7 +906,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc0", "EventName": "INST_RETIRED.ANY_P", - "PublicDescription": "Counts the number of X86 instructions retire= d - an Architectural PerfMon event. Counting continues during hardware inte= rrupts, traps, and inside interrupt handlers. Notes: INST_RETIRED.ANY is co= unted by a designated fixed counter freeing up programmable counters to cou= nt other events. INST_RETIRED.ANY_P is counted by a programmable counter. A= vailable PDIST counters: 0", + "PublicDescription": "Counts the number of X86 instructions retire= d - an Architectural PerfMon event. Counting continues during hardware inte= rrupts, traps, and inside interrupt handlers. Notes: INST_RETIRED.ANY is co= unted by a designated fixed counter freeing up programmable counters to cou= nt other events. INST_RETIRED.ANY_P is counted by a programmable counter. A= vailable PDIST counters: 0,1", "SampleAfterValue": "2000003", "Unit": "cpu_core" }, @@ -915,7 +915,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc0", "EventName": "INST_RETIRED.BR_FUSED", - "PublicDescription": "retired macro-fused uops when there is a bra= nch in the macro-fused pair (the two instructions that got macro-fused coun= t once in this pmon) Available PDIST counters: 0", + "PublicDescription": "retired macro-fused uops when there is a bra= nch in the macro-fused pair (the two instructions that got macro-fused coun= t once in this pmon) Available PDIST counters: 0,1", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -925,7 +925,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc0", "EventName": "INST_RETIRED.MACRO_FUSED", - "PublicDescription": "INST_RETIRED.MACRO_FUSED Available PDIST cou= nters: 0", + "PublicDescription": "INST_RETIRED.MACRO_FUSED Available PDIST cou= nters: 0,1", "SampleAfterValue": "2000003", "UMask": "0x30", "Unit": "cpu_core" @@ -935,7 +935,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc0", "EventName": "INST_RETIRED.NOP", - "PublicDescription": "Counts all retired NOP or ENDBR32/64 or PREF= ETCHIT0/1 instructions Available PDIST counters: 0", + "PublicDescription": "Counts all retired NOP or ENDBR32/64 or PREF= ETCHIT0/1 instructions Available PDIST counters: 0,1", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -954,7 +954,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xc0", "EventName": "INST_RETIRED.REP_ITERATION", - "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent. Available PDIST counters: 0", + "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent. Available PDIST counters: 0,1", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -1227,6 +1227,15 @@ "UMask": "0x88", "Unit": "cpu_core" }, + { + "BriefDescription": "Counts the number of times a load got early b= locked due to preceding store operation with unknown address or unknown dat= a. Excluding in-line (immediate) wakeups", + "Counter": "0,1,2,3,4,5,6,7,8,9", + "EventCode": "0x03", + "EventName": "LD_BLOCKS.STORE_EARLY", + "SampleAfterValue": "100003", + "UMask": "0xa1", + "Unit": "cpu_core" + }, { "BriefDescription": "Counts the number of occurrences a retired lo= ad gets blocked because its address partially overlaps with an older store = (size mismatch) - unknown_sta/bad_forward", "Counter": "0,1,2,3,4,5,6,7", @@ -1451,7 +1460,7 @@ "Counter": "0,1,2,3,4,5,6,7,8,9", "EventCode": "0xe4", "EventName": "MISC_RETIRED.LBR_INSERTS", - "PublicDescription": "LBR record is inserted Available PDIST count= ers: 0", + "PublicDescription": "LBR record is inserted Available PDIST count= ers: 0,1", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/lunarlake/uncore-interconnect.j= son b/tools/perf/pmu-events/arch/x86/lunarlake/uncore-interconnect.json new file mode 100644 index 000000000000..69ef928d57f6 --- /dev/null +++ b/tools/perf/pmu-events/arch/x86/lunarlake/uncore-interconnect.json @@ -0,0 +1,10 @@ +[ + { + "BriefDescription": "This 48-bit fixed counter counts the UCLK cyc= les.", + "Counter": "FIXED", + "EventCode": "0xff", + "EventName": "UNC_CLOCK.SOCKET", + "PerPkg": "1", + "Unit": "SANTA" + } +] diff --git a/tools/perf/pmu-events/arch/x86/lunarlake/uncore-memory.json b/= tools/perf/pmu-events/arch/x86/lunarlake/uncore-memory.json index 7d63580302de..63c4aa2791e4 100644 --- a/tools/perf/pmu-events/arch/x86/lunarlake/uncore-memory.json +++ b/tools/perf/pmu-events/arch/x86/lunarlake/uncore-memory.json @@ -32,5 +32,13 @@ "Experimental": "1", "PerPkg": "1", "Unit": "iMC" + }, + { + "BriefDescription": "Total number of read and write byte transfers= to/from DRAM, in 32B chunk, per DDR channel. Counter increments by 1 after= sending or receiving 32B chunk data.", + "Counter": "0,1,2,3,4", + "EventCode": "0x3C", + "EventName": "UNC_M_TOTAL_DATA", + "PerPkg": "1", + "Unit": "iMC" } ] diff --git a/tools/perf/pmu-events/arch/x86/mapfile.csv b/tools/perf/pmu-ev= ents/arch/x86/mapfile.csv index 2d7a9fa055dd..894a3f485a4a 100644 --- a/tools/perf/pmu-events/arch/x86/mapfile.csv +++ b/tools/perf/pmu-events/arch/x86/mapfile.csv @@ -22,7 +22,7 @@ GenuineIntel-6-3A,v24,ivybridge,core GenuineIntel-6-3E,v24,ivytown,core GenuineIntel-6-2D,v24,jaketown,core GenuineIntel-6-(57|85),v16,knightslanding,core -GenuineIntel-6-BD,v1.14,lunarlake,core +GenuineIntel-6-BD,v1.17,lunarlake,core GenuineIntel-6-(AA|AC|B5),v1.14,meteorlake,core GenuineIntel-6-1[AEF],v4,nehalemep,core GenuineIntel-6-2E,v4,nehalemex,core --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-qk1-f201.google.com (mail-qk1-f201.google.com [209.85.222.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id C9D53320CBF for ; Mon, 18 Aug 2025 19:06:10 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.222.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543982; cv=none; b=s3NxXHQDLRqy/yMkp8Fyffiu2nxCwwV9mULY/l2nhlheZNZAHSI0cM1X0CAR9petfdj76SRH7i6WppV0JN+l9+BE15UNJTd56fm3mQLGkuVV4JpPyZugEngi0zslawgB8I/lSGiNFOyre1VNKKIuXNg9UHJbNXQ/qkKud1IJ8Qc= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543982; c=relaxed/simple; bh=f2Pf/yPB+qVwWL6qqapYZpg9orroNXEYZs4oU8iPIJU=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=X03e1R2VnJIvxcFIilYoAp6GnUPjw7X0uyGuM5pOSjc7XvGrTfeI1BHKRQNdJ1X4ex+Ao2IbJDLCIrU4FL6E0WmDgLs/1SKO4ioB32zSdqGLjv5aCGS2WKTDdpWkk2FzTLSH2w02gXEE5owxOaUhDmkSntp/e1JmXLd/NsnZA5M= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=Mqp/kA8x; arc=none smtp.client-ip=209.85.222.201 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="Mqp/kA8x" Received: by mail-qk1-f201.google.com with SMTP id af79cd13be357-7e8702fb9e1so1267920685a.0 for ; Mon, 18 Aug 2025 12:06:10 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543969; x=1756148769; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=Jaoiyo+TuDbgc1dxi8QIt8Pw8Fm8ix4xZqpIPrOA3Eo=; b=Mqp/kA8xSTiFB8ZXq0mI9Junzo6JmSYW+gAdcZIFKfMWUG9MbeLa/JDlPrNRYM85KI W+7vr7mZr6gr6EiPe7KhATjqbilxjSoz10UM7rZ06aoTmt6Pc5y/ZpI9yceeMYsl9uZv d8lbUuAIgzd9BfdHktWl8TbQYgVMcz34dZW7HX37AwSuVUc4cjNmm0qtPeD6sMTxiGT5 Ad10o54v0Ey7PdUwX3/dT+PSDQI594p4K3ehRsUW1psI27o32etLv8amaM9lJc2sCOG2 UjlU6MJst6OXTrinv1jAtsQxADHVK2vPhqaM2lVxHIYPBxECziIXz2ZbfVUKuiDxRILW SFlQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543969; x=1756148769; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=Jaoiyo+TuDbgc1dxi8QIt8Pw8Fm8ix4xZqpIPrOA3Eo=; b=RW71RCdqMrFFhUk+OsmZKatCo1Ivyu54aglwH/zQT/Hr95v/f9GsfnbaXMKALu5fpF UWBtjBMf82A70sNaE64z/0bgnBD5RYNsQhhk5P6coHGQcmguLjKRGALBo3Q/d/OM97EW u6l37sarlQPC/5RYLjxziJUq1/bALhc2hSvn7RXAid0W1MYBQBy5MXw0KJzgmvDAiss5 bnw9rJR9a+U+3inO4iWhlnIeXgSahrpjbtlVTeam99zEd948es6pzLfABZqwifersS32 t2Kz845jN0CA73X+Jc4OdMWW0plh4ZenjiM15MbWiRzZ7i3q2EO+EmnKDHuogKcIrfyp scZQ== X-Forwarded-Encrypted: i=1; AJvYcCU7Uoy1gQqC30Ce+7yqWS6wThXgYgOxpMHZE5PAe/5h90v395SB22Sfgk7URcS1ooI8FGJFlbWJ+bC7UdU=@vger.kernel.org X-Gm-Message-State: AOJu0Yy21Ymvml+6mZ/upA0R7lwzHNRIqFomPjEWbeCQDNfNsKCjxHPH 4+A7+5u0EXPGSAQSoIUFnbmV4hHuBgYDPRkoLt2I6kOSNfIkz808pfV+dPC+aEevGThMT+PW1Gw WJOGyPPfOVw== X-Google-Smtp-Source: AGHT+IEt/xDMHTrsvFQDeEnUHJ7yv0fAuFlEIGOf3TxuKjkO4wu2EEhtKbLWOVGOipgnkfLq+cqv7emtqoCg X-Received: from qkpc28.prod.google.com ([2002:a05:620a:269c:b0:7e8:8ef1:d85d]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a05:620a:1989:b0:7e8:3152:1ac3 with SMTP id af79cd13be357-7e9f2cbd6a8mr45189885a.41.1755543969300; Mon, 18 Aug 2025 12:06:09 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:10 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-15-irogers@google.com> Subject: [PATCH v2 14/20] perf vendor events: Update meteorlake events/metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update events from v1.14 to v1.16. Update metrics from TMA 5.0 to 5.1. The event updates come from: https://github.com/intel/perfmon/commit/c3e91c6e6b39429c57001d4942667f380ef= e8ea9 Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- .../pmu-events/arch/x86/meteorlake/cache.json | 129 ++++++------- .../arch/x86/meteorlake/floating-point.json | 28 +-- .../arch/x86/meteorlake/frontend.json | 42 ++--- .../arch/x86/meteorlake/memory.json | 15 +- .../arch/x86/meteorlake/mtl-metrics.json | 103 ++++++----- .../pmu-events/arch/x86/meteorlake/other.json | 5 +- .../arch/x86/meteorlake/pipeline.json | 173 ++++++++---------- .../arch/x86/meteorlake/virtual-memory.json | 40 ++-- 9 files changed, 251 insertions(+), 286 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/mapfile.csv b/tools/perf/pmu-ev= ents/arch/x86/mapfile.csv index 894a3f485a4a..f8bf16d30602 100644 --- a/tools/perf/pmu-events/arch/x86/mapfile.csv +++ b/tools/perf/pmu-events/arch/x86/mapfile.csv @@ -23,7 +23,7 @@ GenuineIntel-6-3E,v24,ivytown,core GenuineIntel-6-2D,v24,jaketown,core GenuineIntel-6-(57|85),v16,knightslanding,core GenuineIntel-6-BD,v1.17,lunarlake,core -GenuineIntel-6-(AA|AC|B5),v1.14,meteorlake,core +GenuineIntel-6-(AA|AC|B5),v1.16,meteorlake,core GenuineIntel-6-1[AEF],v4,nehalemep,core GenuineIntel-6-2E,v4,nehalemex,core GenuineIntel-6-CC,v1.00,pantherlake,core diff --git a/tools/perf/pmu-events/arch/x86/meteorlake/cache.json b/tools/p= erf/pmu-events/arch/x86/meteorlake/cache.json index 82b115183924..f1d8db284a02 100644 --- a/tools/perf/pmu-events/arch/x86/meteorlake/cache.json +++ b/tools/perf/pmu-events/arch/x86/meteorlake/cache.json @@ -14,7 +14,6 @@ "Counter": "0,1,2,3", "EventCode": "0x51", "EventName": "L1D.HWPF_MISS", - "PublicDescription": "L1D.HWPF_MISS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x20", "Unit": "cpu_core" @@ -24,7 +23,7 @@ "Counter": "0,1,2,3", "EventCode": "0x51", "EventName": "L1D.REPLACEMENT", - "PublicDescription": "Counts L1D data line replacements including = opportunistic replacements, and replacements that require stall-for-replace= or block-for-replace. Available PDIST counters: 0", + "PublicDescription": "Counts L1D data line replacements including = opportunistic replacements, and replacements that require stall-for-replace= or block-for-replace.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -34,7 +33,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.FB_FULL", - "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -46,7 +45,7 @@ "EdgeDetect": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.FB_FULL_PERIODS", - "PublicDescription": "Counts number of phases a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts number of phases a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -56,7 +55,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.L2_STALLS", - "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D due to lack of L2 resources. Demand requests include cac= heable/uncacheable demand load, store, lock or SW prefetch accesses. Availa= ble PDIST counters: 0", + "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D due to lack of L2 resources. Demand requests include cac= heable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x4", "Unit": "cpu_core" @@ -66,7 +65,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.PENDING", - "PublicDescription": "Counts number of L1D misses that are outstan= ding in each cycle, that is each cycle the number of Fill Buffers (FB) outs= tanding required by Demand Reads. FB either is held by demand loads, or it = is held by non-demand loads and gets hit at least once by demand. The valid= outstanding interval is defined until the FB deallocation by one of the fo= llowing ways: from FB allocation, if FB is allocated by demand from the dem= and Hit FB, if it is allocated by hardware or software prefetch. Note: In t= he L1D, a Demand Read contains cacheable or noncacheable demand loads, incl= uding ones causing cache-line splits and reads due to page walks resulted f= rom any request type. Available PDIST counters: 0", + "PublicDescription": "Counts number of L1D misses that are outstan= ding in each cycle, that is each cycle the number of Fill Buffers (FB) outs= tanding required by Demand Reads. FB either is held by demand loads, or it = is held by non-demand loads and gets hit at least once by demand. The valid= outstanding interval is defined until the FB deallocation by one of the fo= llowing ways: from FB allocation, if FB is allocated by demand from the dem= and Hit FB, if it is allocated by hardware or software prefetch. Note: In t= he L1D, a Demand Read contains cacheable or noncacheable demand loads, incl= uding ones causing cache-line splits and reads due to page walks resulted f= rom any request type.", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -77,7 +76,7 @@ "CounterMask": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.PENDING_CYCLES", - "PublicDescription": "Counts duration of L1D miss outstanding in c= ycles. Available PDIST counters: 0", + "PublicDescription": "Counts duration of L1D miss outstanding in c= ycles.", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -87,7 +86,7 @@ "Counter": "0,1,2,3", "EventCode": "0x25", "EventName": "L2_LINES_IN.ALL", - "PublicDescription": "Counts the number of L2 cache lines filling = the L2. Counting does not cover rejects. Available PDIST counters: 0", + "PublicDescription": "Counts the number of L2 cache lines filling = the L2. Counting does not cover rejects.", "SampleAfterValue": "100003", "UMask": "0x1f", "Unit": "cpu_core" @@ -147,7 +146,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.NON_SILENT", - "PublicDescription": "Counts the number of lines that are evicted = by L2 cache when triggered by an L2 cache fill. Those lines are in Modified= state. Modified lines are written back to L3 Available PDIST counters: 0", + "PublicDescription": "Counts the number of lines that are evicted = by L2 cache when triggered by an L2 cache fill. Those lines are in Modified= state. Modified lines are written back to L3", "SampleAfterValue": "200003", "UMask": "0x2", "Unit": "cpu_core" @@ -167,7 +166,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.SILENT", - "PublicDescription": "Counts the number of lines that are silently= dropped by L2 cache. These lines are typically in Shared or Exclusive stat= e. A non-threaded event. Available PDIST counters: 0", + "PublicDescription": "Counts the number of lines that are silently= dropped by L2 cache. These lines are typically in Shared or Exclusive stat= e. A non-threaded event.", "SampleAfterValue": "200003", "UMask": "0x1", "Unit": "cpu_core" @@ -177,7 +176,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.USELESS_HWPF", - "PublicDescription": "Counts the number of cache lines that have b= een prefetched by the L2 hardware prefetcher but not used by demand access = when evicted from the L2 cache Available PDIST counters: 0", + "PublicDescription": "Counts the number of cache lines that have b= een prefetched by the L2 hardware prefetcher but not used by demand access = when evicted from the L2 cache", "SampleAfterValue": "200003", "UMask": "0x4", "Unit": "cpu_core" @@ -187,7 +186,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.ALL", - "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_RQSTS.REFERENCES] Available PDIST coun= ters: 0", + "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_RQSTS.REFERENCES]", "SampleAfterValue": "200003", "UMask": "0xff", "Unit": "cpu_core" @@ -206,7 +205,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.HIT", - "PublicDescription": "Counts all requests that hit L2 cache. [This= event is alias to L2_RQSTS.HIT] Available PDIST counters: 0", + "PublicDescription": "Counts all requests that hit L2 cache. [This= event is alias to L2_RQSTS.HIT]", "SampleAfterValue": "200003", "UMask": "0xdf", "Unit": "cpu_core" @@ -225,7 +224,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.MISS", - "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_RQSTS.MISS] Available PDIST coun= ters: 0", + "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_RQSTS.MISS]", "SampleAfterValue": "200003", "UMask": "0x3f", "Unit": "cpu_core" @@ -244,7 +243,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_CODE_RD", - "PublicDescription": "Counts the total number of L2 code requests.= Available PDIST counters: 0", + "PublicDescription": "Counts the total number of L2 code requests.= ", "SampleAfterValue": "200003", "UMask": "0xe4", "Unit": "cpu_core" @@ -254,7 +253,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_DATA_RD", - "PublicDescription": "Counts Demand Data Read requests accessing t= he L2 cache. These requests may hit or miss L2 cache. True-miss exclude mis= ses that were merged with ongoing L2 misses. An access is counted once. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts Demand Data Read requests accessing t= he L2 cache. These requests may hit or miss L2 cache. True-miss exclude mis= ses that were merged with ongoing L2 misses. An access is counted once.", "SampleAfterValue": "200003", "UMask": "0xe1", "Unit": "cpu_core" @@ -264,7 +263,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_MISS", - "PublicDescription": "Counts demand requests that miss L2 cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts demand requests that miss L2 cache.", "SampleAfterValue": "200003", "UMask": "0x27", "Unit": "cpu_core" @@ -274,7 +273,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_REFERENCES", - "PublicDescription": "Counts demand requests to L2 cache. Availabl= e PDIST counters: 0", + "PublicDescription": "Counts demand requests to L2 cache.", "SampleAfterValue": "200003", "UMask": "0xe7", "Unit": "cpu_core" @@ -284,7 +283,6 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_HWPF", - "PublicDescription": "L2_RQSTS.ALL_HWPF Available PDIST counters: = 0", "SampleAfterValue": "200003", "UMask": "0xf0", "Unit": "cpu_core" @@ -294,7 +292,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_RFO", - "PublicDescription": "Counts the total number of RFO (read for own= ership) requests to L2 cache. L2 RFO requests include both L1D demand RFO m= isses as well as L1D RFO prefetches. Available PDIST counters: 0", + "PublicDescription": "Counts the total number of RFO (read for own= ership) requests to L2 cache. L2 RFO requests include both L1D demand RFO m= isses as well as L1D RFO prefetches.", "SampleAfterValue": "200003", "UMask": "0xe2", "Unit": "cpu_core" @@ -304,7 +302,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.CODE_RD_HIT", - "PublicDescription": "Counts L2 cache hits when fetching instructi= ons, code reads. Available PDIST counters: 0", + "PublicDescription": "Counts L2 cache hits when fetching instructi= ons, code reads.", "SampleAfterValue": "200003", "UMask": "0xc4", "Unit": "cpu_core" @@ -314,7 +312,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.CODE_RD_MISS", - "PublicDescription": "Counts L2 cache misses when fetching instruc= tions. Available PDIST counters: 0", + "PublicDescription": "Counts L2 cache misses when fetching instruc= tions.", "SampleAfterValue": "200003", "UMask": "0x24", "Unit": "cpu_core" @@ -324,7 +322,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT", - "PublicDescription": "Counts the number of demand Data Read reques= ts initiated by load instructions that hit L2 cache. Available PDIST counte= rs: 0", + "PublicDescription": "Counts the number of demand Data Read reques= ts initiated by load instructions that hit L2 cache.", "SampleAfterValue": "200003", "UMask": "0xc1", "Unit": "cpu_core" @@ -334,7 +332,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.DEMAND_DATA_RD_MISS", - "PublicDescription": "Counts demand Data Read requests with true-m= iss in the L2 cache. True-miss excludes misses that were merged with ongoin= g L2 misses. An access is counted once. Available PDIST counters: 0", + "PublicDescription": "Counts demand Data Read requests with true-m= iss in the L2 cache. True-miss excludes misses that were merged with ongoin= g L2 misses. An access is counted once.", "SampleAfterValue": "200003", "UMask": "0x21", "Unit": "cpu_core" @@ -344,7 +342,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.HIT", - "PublicDescription": "Counts all requests that hit L2 cache. [This= event is alias to L2_REQUEST.HIT] Available PDIST counters: 0", + "PublicDescription": "Counts all requests that hit L2 cache. [This= event is alias to L2_REQUEST.HIT]", "SampleAfterValue": "200003", "UMask": "0xdf", "Unit": "cpu_core" @@ -354,7 +352,6 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.HWPF_MISS", - "PublicDescription": "L2_RQSTS.HWPF_MISS Available PDIST counters:= 0", "SampleAfterValue": "200003", "UMask": "0x30", "Unit": "cpu_core" @@ -364,7 +361,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.MISS", - "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_REQUEST.MISS] Available PDIST co= unters: 0", + "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_REQUEST.MISS]", "SampleAfterValue": "200003", "UMask": "0x3f", "Unit": "cpu_core" @@ -374,7 +371,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.REFERENCES", - "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_REQUEST.ALL] Available PDIST counters:= 0", + "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_REQUEST.ALL]", "SampleAfterValue": "200003", "UMask": "0xff", "Unit": "cpu_core" @@ -384,7 +381,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.RFO_HIT", - "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that hit L2 cache. Available PDIST counters: 0", + "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that hit L2 cache.", "SampleAfterValue": "200003", "UMask": "0xc2", "Unit": "cpu_core" @@ -394,7 +391,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.RFO_MISS", - "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that miss L2 cache. Available PDIST counters: 0", + "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that miss L2 cache.", "SampleAfterValue": "200003", "UMask": "0x22", "Unit": "cpu_core" @@ -404,7 +401,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.SWPF_HIT", - "PublicDescription": "Counts Software prefetch requests that hit t= he L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when = FB is not full. Available PDIST counters: 0", + "PublicDescription": "Counts Software prefetch requests that hit t= he L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when = FB is not full.", "SampleAfterValue": "200003", "UMask": "0xc8", "Unit": "cpu_core" @@ -414,7 +411,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.SWPF_MISS", - "PublicDescription": "Counts Software prefetch requests that miss = the L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when= FB is not full. Available PDIST counters: 0", + "PublicDescription": "Counts Software prefetch requests that miss = the L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when= FB is not full.", "SampleAfterValue": "200003", "UMask": "0x28", "Unit": "cpu_core" @@ -424,7 +421,7 @@ "Counter": "0,1,2,3", "EventCode": "0x23", "EventName": "L2_TRANS.L2_WB", - "PublicDescription": "Counts L2 writebacks that access L2 cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts L2 writebacks that access L2 cache.", "SampleAfterValue": "200003", "UMask": "0x40", "Unit": "cpu_core" @@ -434,7 +431,7 @@ "Counter": "0,1,2,3", "EventCode": "0x42", "EventName": "LOCK_CYCLES.CACHE_LOCK_DURATION", - "PublicDescription": "This event counts the number of cycles when = the L1D is locked. It is a superset of the 0x1 mask (BUS_LOCK_CLOCKS.BUS_LO= CK_DURATION). Available PDIST counters: 0", + "PublicDescription": "This event counts the number of cycles when = the L1D is locked. It is a superset of the 0x1 mask (BUS_LOCK_CLOCKS.BUS_LO= CK_DURATION).", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -454,7 +451,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x2e", "EventName": "LONGEST_LAT_CACHE.MISS", - "PublicDescription": "Counts core-originated cacheable requests th= at miss the L3 cache (Longest Latency cache). Requests include data and cod= e reads, Reads-for-Ownership (RFOs), speculative accesses and hardware pref= etches to the L1 and L2. It does not include hardware prefetches to the L3= , and may not count other types of requests to the L3. Available PDIST coun= ters: 0", + "PublicDescription": "Counts core-originated cacheable requests th= at miss the L3 cache (Longest Latency cache). Requests include data and cod= e reads, Reads-for-Ownership (RFOs), speculative accesses and hardware pref= etches to the L1 and L2. It does not include hardware prefetches to the L3= , and may not count other types of requests to the L3.", "SampleAfterValue": "100003", "UMask": "0x41", "Unit": "cpu_core" @@ -474,7 +471,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x2e", "EventName": "LONGEST_LAT_CACHE.REFERENCE", - "PublicDescription": "Counts core-originated cacheable requests to= the L3 cache (Longest Latency cache). Requests include data and code reads= , Reads-for-Ownership (RFOs), speculative accesses and hardware prefetches = to the L1 and L2. It does not include hardware prefetches to the L3, and m= ay not count other types of requests to the L3. Available PDIST counters: 0= ", + "PublicDescription": "Counts core-originated cacheable requests to= the L3 cache (Longest Latency cache). Requests include data and code reads= , Reads-for-Ownership (RFOs), speculative accesses and hardware prefetches = to the L1 and L2. It does not include hardware prefetches to the L3, and m= ay not count other types of requests to the L3.", "SampleAfterValue": "100003", "UMask": "0x4f", "Unit": "cpu_core" @@ -695,7 +692,7 @@ "Counter": "0,1,2,3", "EventCode": "0x43", "EventName": "MEM_LOAD_COMPLETED.L1_MISS_ANY", - "PublicDescription": "Number of completed demand load requests tha= t missed the L1 data cache including shadow misses (FB hits, merge to an on= going L1D miss) Available PDIST counters: 0", + "PublicDescription": "Number of completed demand load requests tha= t missed the L1 data cache including shadow misses (FB hits, merge to an on= going L1D miss)", "SampleAfterValue": "1000003", "UMask": "0xfd", "Unit": "cpu_core" @@ -947,7 +944,6 @@ "Counter": "0,1,2,3", "EventCode": "0x44", "EventName": "MEM_STORE_RETIRED.L2_HIT", - "PublicDescription": "MEM_STORE_RETIRED.L2_HIT Available PDIST cou= nters: 0", "SampleAfterValue": "200003", "UMask": "0x1", "Unit": "cpu_core" @@ -974,7 +970,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_1024", @@ -986,7 +982,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_128", @@ -998,7 +994,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_16", @@ -1010,7 +1006,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_2048", @@ -1022,7 +1018,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_256", @@ -1034,7 +1030,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_32", @@ -1046,7 +1042,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_4", @@ -1058,7 +1054,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_512", @@ -1070,7 +1066,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_64", @@ -1082,7 +1078,7 @@ }, { "BriefDescription": "Counts the number of tagged load uops retired= that exceed the latency threshold defined in MEC_CR_PEBS_LD_LAT_THRESHOLD = - Only counts with PEBS enabled.", - "Counter": "0,1", + "Counter": "0,1,2,3,4,5,6,7", "Data_LA": "1", "EventCode": "0xd0", "EventName": "MEM_UOPS_RETIRED.LOAD_LATENCY_GT_8", @@ -1177,7 +1173,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe5", "EventName": "MEM_UOP_RETIRED.ANY", - "PublicDescription": "Number of retired micro-operations (uops) fo= r load or store memory accesses Available PDIST counters: 0", + "PublicDescription": "Number of retired micro-operations (uops) fo= r load or store memory accesses", "SampleAfterValue": "1000003", "UMask": "0x3", "Unit": "cpu_core" @@ -1403,7 +1399,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.ALL_REQUESTS", - "PublicDescription": "Counts memory transactions reached the super= queue including requests initiated by the core, all L3 prefetches, page wa= lks, etc.. Available PDIST counters: 0", + "PublicDescription": "Counts memory transactions reached the super= queue including requests initiated by the core, all L3 prefetches, page wa= lks, etc..", "SampleAfterValue": "100003", "UMask": "0x80", "Unit": "cpu_core" @@ -1413,7 +1409,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DATA_RD", - "PublicDescription": "Counts the demand and prefetch data reads. A= ll Core Data Reads include cacheable 'Demands' and L2 prefetchers (not L3 p= refetchers). Counting also covers reads due to page walks resulted from any= request type. Available PDIST counters: 0", + "PublicDescription": "Counts the demand and prefetch data reads. A= ll Core Data Reads include cacheable 'Demands' and L2 prefetchers (not L3 p= refetchers). Counting also covers reads due to page walks resulted from any= request type.", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -1423,7 +1419,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_CODE_RD", - "PublicDescription": "Counts both cacheable and Non-Cacheable code= read requests. Available PDIST counters: 0", + "PublicDescription": "Counts both cacheable and Non-Cacheable code= read requests.", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -1433,7 +1429,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_DATA_RD", - "PublicDescription": "Counts the Demand Data Read requests sent to= uncore. Use it in conjunction with OFFCORE_REQUESTS_OUTSTANDING to determi= ne average latency in the uncore. Available PDIST counters: 0", + "PublicDescription": "Counts the Demand Data Read requests sent to= uncore. Use it in conjunction with OFFCORE_REQUESTS_OUTSTANDING to determi= ne average latency in the uncore.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -1443,7 +1439,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_RFO", - "PublicDescription": "Counts the demand RFO (read for ownership) r= equests including regular RFOs, locks, ItoM. Available PDIST counters: 0", + "PublicDescription": "Counts the demand RFO (read for ownership) r= equests including regular RFOs, locks, ItoM.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -1454,7 +1450,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD", - "PublicDescription": "Counts cycles when offcore outstanding cache= able Core Data Read transactions are present in the super queue. A transact= ion is considered to be in the Offcore outstanding state between L2 miss an= d transaction completion sent to requestor (SQ de-allocation). See correspo= nding Umask under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when offcore outstanding cache= able Core Data Read transactions are present in the super queue. A transact= ion is considered to be in the Offcore outstanding state between L2 miss an= d transaction completion sent to requestor (SQ de-allocation). See correspo= nding Umask under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -1465,7 +1461,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_CODE= _RD", - "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1476,7 +1472,6 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA= _RD", - "PublicDescription": "Cycles where at least 1 outstanding demand d= ata read request is pending. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1487,7 +1482,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO", - "PublicDescription": "Counts the number of offcore outstanding dem= and rfo Reads transactions in the super queue every cycle. The 'Offcore out= standing' state of the transaction lasts from the L2 miss until the sending= transaction completion to requestor (SQ deallocation). See the correspondi= ng Umask under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding dem= and rfo Reads transactions in the super queue every cycle. The 'Offcore out= standing' state of the transaction lasts from the L2 miss until the sending= transaction completion to requestor (SQ deallocation). See the correspondi= ng Umask under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x4", "Unit": "cpu_core" @@ -1497,7 +1492,6 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DATA_RD", - "PublicDescription": "OFFCORE_REQUESTS_OUTSTANDING.DATA_RD Availab= le PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -1507,7 +1501,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_CODE_RD", - "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1517,7 +1511,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD", - "PublicDescription": "For every cycle, increments by the number of= outstanding demand data read requests pending. Requests are considered o= utstanding from the time they miss the core's L2 cache until the transactio= n completion message is sent to the requestor. Available PDIST counters: 0", + "PublicDescription": "For every cycle, increments by the number of= outstanding demand data read requests pending. Requests are considered o= utstanding from the time they miss the core's L2 cache until the transactio= n completion message is sent to the requestor.", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1527,7 +1521,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_RFO", - "PublicDescription": "Counts the number of off-core outstanding re= ad-for-ownership (RFO) store transactions every cycle. An RFO transaction i= s considered to be in the Off-core outstanding state between L2 cache miss = and transaction completion. Available PDIST counters: 0", + "PublicDescription": "Counts the number of off-core outstanding re= ad-for-ownership (RFO) store transactions every cycle. An RFO transaction i= s considered to be in the Off-core outstanding state between L2 cache miss = and transaction completion.", "SampleAfterValue": "1000003", "UMask": "0x4", "Unit": "cpu_core" @@ -1537,7 +1531,7 @@ "Counter": "0,1,2,3", "EventCode": "0x2c", "EventName": "SQ_MISC.BUS_LOCK", - "PublicDescription": "Counts the more expensive bus lock needed to= enforce cache coherency for certain memory accesses that need to be done a= tomically. Can be created by issuing an atomic instruction (via the LOCK p= refix) which causes a cache line split or accesses uncacheable memory. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts the more expensive bus lock needed to= enforce cache coherency for certain memory accesses that need to be done a= tomically. Can be created by issuing an atomic instruction (via the LOCK p= refix) which causes a cache line split or accesses uncacheable memory.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -1547,7 +1541,6 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.ANY", - "PublicDescription": "Counts the number of PREFETCHNTA, PREFETCHW,= PREFETCHT0, PREFETCHT1 or PREFETCHT2 instructions executed. Available PDIS= T counters: 0", "SampleAfterValue": "100003", "UMask": "0xf", "Unit": "cpu_core" @@ -1557,7 +1550,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.NTA", - "PublicDescription": "Counts the number of PREFETCHNTA instruction= s executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHNTA instruction= s executed.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -1567,7 +1560,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.PREFETCHW", - "PublicDescription": "Counts the number of PREFETCHW instructions = executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHW instructions = executed.", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -1577,7 +1570,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.T0", - "PublicDescription": "Counts the number of PREFETCHT0 instructions= executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHT0 instructions= executed.", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -1587,7 +1580,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.T1_T2", - "PublicDescription": "Counts the number of PREFETCHT1 or PREFETCHT= 2 instructions executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHT1 or PREFETCHT= 2 instructions executed.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/meteorlake/floating-point.json = b/tools/perf/pmu-events/arch/x86/meteorlake/floating-point.json index ae9778aa755b..28dc5e06ee31 100644 --- a/tools/perf/pmu-events/arch/x86/meteorlake/floating-point.json +++ b/tools/perf/pmu-events/arch/x86/meteorlake/floating-point.json @@ -15,7 +15,6 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.FPDIV_ACTIVE", - "PublicDescription": "This event counts the cycles the floating po= int divider is busy. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -25,7 +24,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.FP", - "PublicDescription": "Counts all microcode Floating Point assists.= Available PDIST counters: 0", + "PublicDescription": "Counts all microcode Floating Point assists.= ", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -35,7 +34,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.SSE_AVX_MIX", - "PublicDescription": "ASSISTS.SSE_AVX_MIX Available PDIST counters= : 0", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -45,7 +43,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_0", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_0 [This event is al= ias to FP_ARITH_DISPATCHED.V0] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -55,7 +52,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_1", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_1 [This event is al= ias to FP_ARITH_DISPATCHED.V1] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -65,7 +61,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_5", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_5 [This event is al= ias to FP_ARITH_DISPATCHED.V2] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -75,7 +70,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V0", - "PublicDescription": "FP_ARITH_DISPATCHED.V0 [This event is alias = to FP_ARITH_DISPATCHED.PORT_0] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -85,7 +79,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V1", - "PublicDescription": "FP_ARITH_DISPATCHED.V1 [This event is alias = to FP_ARITH_DISPATCHED.PORT_1] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -95,7 +88,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V2", - "PublicDescription": "FP_ARITH_DISPATCHED.V2 [This event is alias = to FP_ARITH_DISPATCHED.PORT_5] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -105,7 +97,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 2 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as the= y perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR re= gister need to be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 2 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as the= y perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR re= gister need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -115,7 +107,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events. Available PDIST co= unters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -125,7 +117,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 = calculations per element. The DAZ and FTZ flags in the MXCSR register need = to be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 = calculations per element. The DAZ and FTZ flags in the MXCSR register need = to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -135,7 +127,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events. Available PDIST co= unters: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -145,7 +137,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.4_FLOPS", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision and 256-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 2 or/and 4 computation operations, one for each element. = Applies to SSE* and AVX* packed single precision floating-point and packed= double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL= DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB ins= tructions count twice as they perform 2 calculations per element. The DAZ a= nd FTZ flags in the MXCSR register need to be set when using these events. = Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision and 256-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 2 or/and 4 computation operations, one for each element. = Applies to SSE* and AVX* packed single precision floating-point and packed= double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL= DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB ins= tructions count twice as they perform 2 calculations per element. The DAZ a= nd FTZ flags in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x18", "Unit": "cpu_core" @@ -155,7 +147,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR", - "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision and double precision floating-point instructions retired; some = instructions will count twice as noted below. Each count represents 1 comp= utational operation. Applies to SSE* and AVX* scalar single precision float= ing-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB= . FM(N)ADD/SUB instructions count twice as they perform 2 calculations per= element. The DAZ and FTZ flags in the MXCSR register need to be set when u= sing these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision and double precision floating-point instructions retired; some = instructions will count twice as noted below. Each count represents 1 comp= utational operation. Applies to SSE* and AVX* scalar single precision float= ing-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB= . FM(N)ADD/SUB instructions count twice as they perform 2 calculations per= element. The DAZ and FTZ flags in the MXCSR register need to be set when u= sing these events.", "SampleAfterValue": "1000003", "UMask": "0x3", "Unit": "cpu_core" @@ -165,7 +157,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational scalar doubl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar double precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions co= unt twice as they perform 2 calculations per element. The DAZ and FTZ flags= in the MXCSR register need to be set when using these events. Available PD= IST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar doubl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar double precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions co= unt twice as they perform 2 calculations per element. The DAZ and FTZ flags= in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -175,7 +167,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE", - "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar single precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instr= uctions count twice as they perform 2 calculations per element. The DAZ and= FTZ flags in the MXCSR register need to be set when using these events. Av= ailable PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar single precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instr= uctions count twice as they perform 2 calculations per element. The DAZ and= FTZ flags in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -185,7 +177,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.VECTOR", - "PublicDescription": "Number of any Vector retired FP arithmetic i= nstructions. The DAZ and FTZ flags in the MXCSR register need to be set wh= en using these events. Available PDIST counters: 0", + "PublicDescription": "Number of any Vector retired FP arithmetic i= nstructions. The DAZ and FTZ flags in the MXCSR register need to be set wh= en using these events.", "SampleAfterValue": "1000003", "UMask": "0xfc", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/meteorlake/frontend.json b/tool= s/perf/pmu-events/arch/x86/meteorlake/frontend.json index 82727022efb6..6484834b1127 100644 --- a/tools/perf/pmu-events/arch/x86/meteorlake/frontend.json +++ b/tools/perf/pmu-events/arch/x86/meteorlake/frontend.json @@ -14,7 +14,7 @@ "Counter": "0,1,2,3", "EventCode": "0x60", "EventName": "BACLEARS.ANY", - "PublicDescription": "Number of times the front-end is resteered w= hen it finds a branch instruction in a fetch line. This is called Unknown B= ranch which occurs for the first time a branch instruction is fetched or wh= en the branch is not tracked by the BPU (Branch Prediction Unit) anymore. A= vailable PDIST counters: 0", + "PublicDescription": "Number of times the front-end is resteered w= hen it finds a branch instruction in a fetch line. This is called Unknown B= ranch which occurs for the first time a branch instruction is fetched or wh= en the branch is not tracked by the BPU (Branch Prediction Unit) anymore.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -24,7 +24,7 @@ "Counter": "0,1,2,3", "EventCode": "0x87", "EventName": "DECODE.LCP", - "PublicDescription": "Counts cycles that the Instruction Length de= coder (ILD) stalls occurred due to dynamically changing prefix length of th= e decoded instruction (by operand size prefix instruction 0x66, address siz= e prefix instruction 0x67 or REX.W for Intel64). Count is proportional to t= he number of prefixes in a 16B-line. This may result in a three-cycle penal= ty for each LCP (Length changing prefix) in a 16-byte chunk. Available PDIS= T counters: 0", + "PublicDescription": "Counts cycles that the Instruction Length de= coder (ILD) stalls occurred due to dynamically changing prefix length of th= e decoded instruction (by operand size prefix instruction 0x66, address siz= e prefix instruction 0x67 or REX.W for Intel64). Count is proportional to t= he number of prefixes in a 16B-line. This may result in a three-cycle penal= ty for each LCP (Length changing prefix) in a 16-byte chunk.", "SampleAfterValue": "500009", "UMask": "0x1", "Unit": "cpu_core" @@ -34,7 +34,6 @@ "Counter": "0,1,2,3", "EventCode": "0x87", "EventName": "DECODE.MS_BUSY", - "PublicDescription": "Cycles the Microcode Sequencer is busy. Avai= lable PDIST counters: 0", "SampleAfterValue": "500009", "UMask": "0x2", "Unit": "cpu_core" @@ -44,7 +43,7 @@ "Counter": "0,1,2,3", "EventCode": "0x61", "EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES", - "PublicDescription": "Decode Stream Buffer (DSB) is a Uop-cache th= at holds translations of previously fetched instructions that were decoded = by the legacy x86 decode pipeline (MITE). This event counts fetch penalty c= ycles when a transition occurs from DSB to MITE. Available PDIST counters: = 0", + "PublicDescription": "Decode Stream Buffer (DSB) is a Uop-cache th= at holds translations of previously fetched instructions that were decoded = by the legacy x86 decode pipeline (MITE). This event counts fetch penalty c= ycles when a transition occurs from DSB to MITE.", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -405,7 +404,7 @@ "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE_DATA.STALLS", - "PublicDescription": "Counts cycles where a code line fetch is sta= lled due to an L1 instruction cache miss. The decode pipeline works at a 32= Byte granularity. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where a code line fetch is sta= lled due to an L1 instruction cache miss. The decode pipeline works at a 32= Byte granularity.", "SampleAfterValue": "500009", "UMask": "0x4", "Unit": "cpu_core" @@ -417,7 +416,6 @@ "EdgeDetect": "1", "EventCode": "0x80", "EventName": "ICACHE_DATA.STALL_PERIODS", - "PublicDescription": "ICACHE_DATA.STALL_PERIODS Available PDIST co= unters: 0", "SampleAfterValue": "500009", "UMask": "0x4", "Unit": "cpu_core" @@ -427,7 +425,7 @@ "Counter": "0,1,2,3", "EventCode": "0x83", "EventName": "ICACHE_TAG.STALLS", - "PublicDescription": "Counts cycles where a code fetch is stalled = due to L1 instruction cache tag miss. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where a code fetch is stalled = due to L1 instruction cache tag miss.", "SampleAfterValue": "200003", "UMask": "0x4", "Unit": "cpu_core" @@ -438,7 +436,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES_ANY", - "PublicDescription": "Counts the number of cycles uops were delive= red to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) p= ath. Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles uops were delive= red to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) p= ath.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -449,7 +447,7 @@ "CounterMask": "6", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES_OK", - "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the D= SB (Decode Stream Buffer) path. Count includes uops that may 'bypass' the I= DQ. Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the D= SB (Decode Stream Buffer) path. Count includes uops that may 'bypass' the I= DQ.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -459,7 +457,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.DSB_UOPS", - "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Availab= le PDIST counters: 0", + "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -470,7 +468,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES_ANY", - "PublicDescription": "Counts the number of cycles uops were delive= red to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipe= line) path. During these cycles uops are not being delivered from the Decod= e Stream Buffer (DSB). Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles uops were delive= red to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipe= line) path. During these cycles uops are not being delivered from the Decod= e Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -481,7 +479,7 @@ "CounterMask": "6", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES_OK", - "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the M= ITE (legacy decode pipeline) path. During these cycles uops are not being d= elivered from the Decode Stream Buffer (DSB). Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the M= ITE (legacy decode pipeline) path. During these cycles uops are not being d= elivered from the Decode Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -491,7 +489,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_UOPS", - "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the MITE path. This also means that uops are= not being delivered from the Decode Stream Buffer (DSB). Available PDIST c= ounters: 0", + "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the MITE path. This also means that uops are= not being delivered from the Decode Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -502,7 +500,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_CYCLES_ANY", - "PublicDescription": "Counts cycles during which uops are being de= livered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS= ) is busy. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts cycles during which uops are being de= livered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS= ) is busy. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.", "SampleAfterValue": "2000003", "UMask": "0x20", "Unit": "cpu_core" @@ -514,7 +512,7 @@ "EdgeDetect": "1", "EventCode": "0x79", "EventName": "IDQ.MS_SWITCHES", - "PublicDescription": "Number of switches from DSB (Decode Stream B= uffer) or MITE (legacy decode pipeline) to the Microcode Sequencer. Availab= le PDIST counters: 0", + "PublicDescription": "Number of switches from DSB (Decode Stream B= uffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -524,7 +522,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_UOPS", - "PublicDescription": "Counts the number of uops initiated by MITE = or Decode Stream Buffer (DSB) and delivered to Instruction Decode Queue (ID= Q) while the Microcode Sequencer (MS) is busy. Counting includes uops that = may 'bypass' the IDQ. Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops initiated by MITE = or Decode Stream Buffer (DSB) and delivered to Instruction Decode Queue (ID= Q) while the Microcode Sequencer (MS) is busy. Counting includes uops that = may 'bypass' the IDQ.", "SampleAfterValue": "1000003", "UMask": "0x20", "Unit": "cpu_core" @@ -534,7 +532,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CORE", - "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that when no operation was delivered to the back-end pipeline due= to instruction fetch limitations when the back-end could have accepted mor= e operations. Common examples include instruction cache misses or x86 instr= uction decode limitations. The count may be distributed among unhalted logi= cal processors (hyper-threads) who share the same physical core, in process= ors that support Intel Hyper-Threading Technology. Software can use this ev= ent as the numerator for the Frontend Bound metric (or top-level category) = of the Top-down Microarchitecture Analysis method. Available PDIST counters= : 0", + "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that when no operation was delivered to the back-end pipeline due= to instruction fetch limitations when the back-end could have accepted mor= e operations. Common examples include instruction cache misses or x86 instr= uction decode limitations. The count may be distributed among unhalted logi= cal processors (hyper-threads) who share the same physical core, in process= ors that support Intel Hyper-Threading Technology. Software can use this ev= ent as the numerator for the Frontend Bound metric (or top-level category) = of the Top-down Microarchitecture Analysis method.", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -545,7 +543,7 @@ "CounterMask": "6", "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CYCLES_0_UOPS_DELIV.CORE", - "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CYCLES= _0_UOPS_DELIV.CORE] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CYCLES= _0_UOPS_DELIV.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -557,7 +555,7 @@ "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CYCLES_FE_WAS_OK", "Invert": "1", - "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_UOPS_N= OT_DELIVERED.CYCLES_FE_WAS_OK] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_UOPS_N= OT_DELIVERED.CYCLES_FE_WAS_OK]", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -567,7 +565,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CORE", - "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle.", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -578,7 +576,7 @@ "CounterMask": "6", "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE", - "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_BUBBLES.CYCLES_0_UOPS_DEL= IV.CORE] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_BUBBLES.CYCLES_0_UOPS_DEL= IV.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -590,7 +588,7 @@ "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK", "Invert": "1", - "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_BUBBLE= S.CYCLES_FE_WAS_OK] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_BUBBLE= S.CYCLES_FE_WAS_OK]", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/meteorlake/memory.json b/tools/= perf/pmu-events/arch/x86/meteorlake/memory.json index 17b94f810d5a..f0cbeda4d5ca 100644 --- a/tools/perf/pmu-events/arch/x86/meteorlake/memory.json +++ b/tools/perf/pmu-events/arch/x86/meteorlake/memory.json @@ -5,7 +5,6 @@ "CounterMask": "2", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L3_MISS", - "PublicDescription": "Cycles while L3 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -16,7 +15,6 @@ "CounterMask": "6", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L3_MISS", - "PublicDescription": "Execution stalls while L3 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x6", "Unit": "cpu_core" @@ -90,7 +88,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.MEMORY_ORDERING", - "PublicDescription": "Counts the number of Machine Clears detected= dye to memory ordering. Memory Ordering Machine Clears may apply when a me= mory read may not conform to the memory ordering rules of the x86 architect= ure Available PDIST counters: 0", + "PublicDescription": "Counts the number of Machine Clears detected= dye to memory ordering. Memory Ordering Machine Clears may apply when a me= mory read may not conform to the memory ordering rules of the x86 architect= ure", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -101,7 +99,6 @@ "CounterMask": "2", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.CYCLES_L1D_MISS", - "PublicDescription": "Cycles while L1 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -112,7 +109,6 @@ "CounterMask": "3", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L1D_MISS", - "PublicDescription": "Execution stalls while L1 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x3", "Unit": "cpu_core" @@ -123,7 +119,7 @@ "CounterMask": "5", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L2_MISS", - "PublicDescription": "Execution stalls while L2 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock). Available PDIST counters: 0", + "PublicDescription": "Execution stalls while L2 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock).", "SampleAfterValue": "1000003", "UMask": "0x5", "Unit": "cpu_core" @@ -134,7 +130,7 @@ "CounterMask": "9", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L3_MISS", - "PublicDescription": "Execution stalls while L3 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock). Available PDIST counters: 0", + "PublicDescription": "Execution stalls while L3 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock).", "SampleAfterValue": "1000003", "UMask": "0x9", "Unit": "cpu_core" @@ -411,7 +407,6 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.L3_MISS_DEMAND_DATA_RD", - "PublicDescription": "Counts demand data read requests that miss t= he L3 cache. Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -422,7 +417,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_L3_MISS_DEM= AND_DATA_RD", - "PublicDescription": "Cycles with at least 1 Demand Data Read requ= ests who miss L3 cache in the superQ. Available PDIST counters: 0", + "PublicDescription": "Cycles with at least 1 Demand Data Read requ= ests who miss L3 cache in the superQ.", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -432,7 +427,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.L3_MISS_DEMAND_DATA_RD", - "PublicDescription": "For every cycle, increments by the number of= demand data read requests pending that are known to have missed the L3 cac= he. Note that this does not capture all elapsed cycles while requests are = outstanding - only cycles from when the requests were known by the requesti= ng core to have missed the L3 cache. Available PDIST counters: 0", + "PublicDescription": "For every cycle, increments by the number of= demand data read requests pending that are known to have missed the L3 cac= he. Note that this does not capture all elapsed cycles while requests are = outstanding - only cycles from when the requests were known by the requesti= ng core to have missed the L3 cache.", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/meteorlake/mtl-metrics.json b/t= ools/perf/pmu-events/arch/x86/meteorlake/mtl-metrics.json index 0088be169f9b..948c16a1f95b 100644 --- a/tools/perf/pmu-events/arch/x86/meteorlake/mtl-metrics.json +++ b/tools/perf/pmu-events/arch/x86/meteorlake/mtl-metrics.json @@ -1,56 +1,56 @@ [ { "BriefDescription": "C10 residency percent per package", - "MetricExpr": "cstate_pkg@c10\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c10\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C10_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C8 residency percent per package", - "MetricExpr": "cstate_pkg@c8\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c8\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C8_Pkg_Residency", "ScaleUnit": "100%" @@ -541,7 +541,7 @@ }, { "BriefDescription": "Average CPU Utilization", - "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.REF_TSC@ / TSC", + "MetricExpr": "cpu_atom@CPU_CLK_UNHALTED.REF_TSC@ / msr@tsc\\,cpu= =3Dcpu_atom@", "MetricName": "tma_info_system_cpu_utilization", "Unit": "cpu_atom" }, @@ -748,7 +748,7 @@ { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-be\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-be\\-bound@ / (cpu_core@topdown\= \-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retirin= g@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BvOB;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", @@ -786,12 +786,21 @@ "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)", "Unit": "cpu_core" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: ", + "Unit": "cpu_core" + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)= ))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full", "Unit": "cpu_core" }, @@ -799,23 +808,14 @@ "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependen= cy + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_= bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma= _l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_dtlb_load + tma_fb= _full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tm= a_store_fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tm= a_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_split_l= oads / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_= latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_s= tore_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_sharin= g + tma_split_stores + tma_store_latency + tma_streaming_stores)) + tma_mem= ory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_boun= d + tma_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store= + tma_false_sharing + tma_split_stores + tma_store_latency + tma_streaming= _stores)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency", "Unit": "cpu_core" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: ", - "Unit": "cpu_core" - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - c= pu_core@INST_RETIRED.REP_ITERATION@ / cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D= 1@) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cl= ears_resteers + tma_mispredicts_resteers * tma_other_mispredicts / tma_bran= ch_mispredicts) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unk= nown_branches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_miss= es + tma_itlb_misses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * t= ma_ms / (tma_dsb + tma_lsd + tma_mite + tma_ms))) - tma_bottleneck_big_code= ", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - c= pu_core@INST_RETIRED.REP_ITERATION@ / cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D= 1@) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cl= ears_resteers + tma_mispredicts_resteers * tma_other_mispredicts / tma_bran= ch_mispredicts) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unk= nown_branches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_miss= es + tma_itlb_misses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottlene= ck_big_code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20", @@ -823,7 +823,7 @@ }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * ((1 - cpu_core@INST_RETIRED.REP_ITERATION@ / = cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_restee= rs * tma_other_mispredicts / tma_branch_mispredicts) / (tma_clears_resteers= + tma_mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers= + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_m= s_switches) + tma_fetch_bandwidth * tma_ms / (tma_dsb + tma_lsd + tma_mite = + tma_ms)) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_bra= nch_mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_n= ukes / tma_other_nukes + tma_core_bound * (tma_serializing_operation + cpu_= core@RS.EMPTY_RESOURCE@ / tma_info_thread_clks * tma_ports_utilized_0) / (t= ma_divider + tma_ports_utilization + tma_serializing_operation) + tma_micro= code_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) * (t= ma_assists / tma_microcode_sequencer) * tma_heavy_operations)", + "MetricExpr": "100 * ((1 - cpu_core@INST_RETIRED.REP_ITERATION@ / = cpu_core@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_restee= rs * tma_other_mispredicts / tma_branch_mispredicts) / (tma_clears_resteers= + tma_mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers= + tma_dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_m= s_switches) + tma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredict= s / tma_branch_mispredicts * tma_branch_mispredicts + tma_machine_clears * = tma_other_nukes / tma_other_nukes + tma_core_bound * (tma_serializing_opera= tion + cpu_core@RS.EMPTY_RESOURCE@ / tma_info_thread_clks * tma_ports_utili= zed_0) / (tma_divider + tma_ports_utilization + tma_serializing_operation) = + tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequ= encer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -859,7 +859,7 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -876,7 +876,7 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU has wasted due to Branch Misprediction", - "MetricExpr": "cpu_core@topdown\\-br\\-mispredict@ / (cpu_core@top= down\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-re= tiring@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-br\\-mispredict@ / (cpu_core@top= down\\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-re= tiring@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BadSpec;BrMispredicts;BvMP;TmaL2;TopdownL2;tma_L2_= group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_specula= tion > 0.15", @@ -1007,7 +1007,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to contested acces= ses", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS@ * min(= cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_MISS@R, 24 * tma_info_system_core_fre= quency) + cpu_core@MEM_LOAD_L3_HIT_RETIRED.XSNP_FWD@ * min(cpu_core@MEM_LOA= D_L3_HIT_RETIRED.XSNP_FWD@R, 25 * tma_info_system_core_frequency) * (cpu_co= re@OCR.DEMAND_DATA_RD.L3_HIT.SNOOP_HITM@ / (cpu_core@OCR.DEMAND_DATA_RD.L3_= HIT.SNOOP_HITM@ + cpu_core@OCR.DEMAND_DATA_RD.L3_HIT.SNOOP_HIT_WITH_FWD@)))= * (1 + cpu_core@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MI= SS@ / 2) / tma_info_thread_clks", "MetricGroup": "BvMS;DataSharing;LockCont;Offcore;Snoop;TopdownL4;= tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", @@ -1124,7 +1123,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -1253,7 +1252,7 @@ }, { "BriefDescription": "This metric represents fraction of slots wher= e the CPU was retiring heavy-weight operations -- instructions that require= two or more uops or micro-coded sequences", - "MetricExpr": "cpu_core@topdown\\-heavy\\-ops@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-heavy\\-ops@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Retire;TmaL2;TopdownL2;tma_L2_group;tma_retiring_g= roup", "MetricName": "tma_heavy_operations", "MetricThreshold": "tma_heavy_operations > 0.1", @@ -1923,7 +1922,7 @@ "Unit": "cpu_core" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "cpu_core@UOPS_EXECUTED.THREAD@ / (cpu_core@UOPS_EXE= CUTED.CORE_CYCLES_GE_1@ / 2 if #SMT_on else cpu_core@UOPS_EXECUTED.THREAD\\= ,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute", @@ -1984,7 +1983,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc\\,cpu= =3Dcpu_core@ / 1e9 / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency", "Unit": "cpu_core" @@ -1998,7 +1997,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.REF_TSC@ / TSC", + "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.REF_TSC@ / msr@tsc\\,cpu= =3Dcpu_core@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized", "Unit": "cpu_core" @@ -2008,7 +2007,7 @@ "MetricExpr": "64 * (UNC_HAC_ARB_TRK_REQUESTS.ALL + UNC_HAC_ARB_CO= H_TRK_REQUESTS.ALL) / 1e9 / tma_info_system_time", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full", + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full", "Unit": "cpu_core" }, { @@ -2094,6 +2093,13 @@ "MetricName": "tma_info_system_turbo_utilization", "Unit": "cpu_core" }, + { + "BriefDescription": "Measured Average Uncore Frequency for the SoC= [GHz]", + "MetricExpr": "tma_info_system_socket_clks / 1e9 / tma_info_system= _time", + "MetricGroup": "SoC", + "MetricName": "tma_info_system_uncore_frequency", + "Unit": "cpu_core" + }, { "BriefDescription": "Per-Logical Processor actual clocks when the = Logical Processor is active.", "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.THREAD@", @@ -2213,12 +2219,12 @@ "Unit": "cpu_core" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", "MetricExpr": "min(2 * (cpu_core@MEM_INST_RETIRED.ALL_LOADS@ - cpu= _core@MEM_LOAD_RETIRED.FB_HIT@ - cpu_core@MEM_LOAD_RETIRED.L1_MISS@) * 20 /= 100, max(cpu_core@CYCLE_ACTIVITY.CYCLES_MEM_ANY@ - cpu_core@MEMORY_ACTIVIT= Y.CYCLES_L1D_MISS@, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2234,7 +2240,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles wit= h demand load accesses that hit the L2 cache under unloaded scenarios (poss= ibly L2 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu_core@MEM_LOAD_RETIRED.L2_HIT@ * min(cpu_core@ME= M_LOAD_RETIRED.L2_HIT@R, 3 * tma_info_system_core_frequency) * (1 + cpu_cor= e@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2) / tma_= info_thread_clks", "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_l2_bound_grou= p", "MetricName": "tma_l2_hit_latency", @@ -2255,12 +2260,11 @@ }, { "BriefDescription": "This metric estimates fraction of cycles with= demand load accesses that hit the L3 cache under unloaded scenarios (possi= bly L3 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu_core@MEM_LOAD_RETIRED.L3_HIT@ * min(cpu_core@ME= M_LOAD_RETIRED.L3_HIT@R, 9 * tma_info_system_core_frequency) * (1 + cpu_cor= e@MEM_LOAD_RETIRED.FB_HIT@ / cpu_core@MEM_LOAD_RETIRED.L1_MISS@ / 2) / tma_= info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2342,6 +2346,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU spent handling cache misses due to lock operations", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu_core@MEM_INST_RETIRED.LOCK_LOADS@ * cpu_core@ME= M_INST_RETIRED.LOCK_LOADS@R / tma_info_thread_clks", "MetricGroup": "LockCont;Offcore;TopdownL4;tma_L4_group;tma_issueR= FO;tma_l1_bound_group", "MetricName": "tma_lock_latency", @@ -2377,7 +2382,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%", "Unit": "cpu_core" }, @@ -2387,13 +2392,13 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%", "Unit": "cpu_core" }, { "BriefDescription": "This metric represents fraction of slots the = Memory subsystem within the Backend was a bottleneck", - "MetricExpr": "cpu_core@topdown\\-mem\\-bound@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-mem\\-bound@ / (cpu_core@topdown= \\-fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiri= ng@ + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "Backend;TmaL2;TopdownL2;tma_L2_group;tma_backend_b= ound_group", "MetricName": "tma_memory_bound", "MetricThreshold": "tma_memory_bound > 0.2 & tma_backend_bound > 0= .2", @@ -2404,7 +2409,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to LFENCE Instructions.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * cpu_core@MISC2_RETIRED.LFENCE@ / tma_info_thre= ad_clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_memory_fence", @@ -2463,7 +2467,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "max(cpu_core@IDQ.MS_CYCLES_ANY@, cpu_core@UOPS_RETI= RED.MS\\,cmask\\=3D1@ / (cpu_core@UOPS_RETIRED.SLOTS@ / cpu_core@UOPS_ISSUE= D.ANY@)) / tma_info_core_core_clks / 2", + "MetricExpr": "max(cpu_core@IDQ.MS_CYCLES_ANY@, cpu_core@UOPS_RETI= RED.MS\\,cmask\\=3D1@ / (cpu_core@UOPS_RETIRED.SLOTS@ / cpu_core@UOPS_ISSUE= D.ANY@)) / tma_info_core_core_clks / 2.4", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -2502,6 +2506,7 @@ }, { "BriefDescription": "This metric represents the remaining light uo= ps fraction the CPU has executed - remaining means not covered by other sib= ling nodes", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_light_operations - (tma_fp_arith + tma_i= nt_operations + tma_memory_operations + tma_fused_instructions + tma_non_fu= sed_branches))", "MetricGroup": "Pipeline;TopdownL3;tma_L3_group;tma_light_operatio= ns_group", "MetricName": "tma_other_light_ops", @@ -2570,6 +2575,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "((tma_ports_utilized_0 * tma_info_thread_clks + (cp= u_core@EXE_ACTIVITY.1_PORTS_UTIL@ + tma_retiring * cpu_core@EXE_ACTIVITY.2_= 3_PORTS_UTIL@)) / tma_info_thread_clks if cpu_core@ARITH.DIV_ACTIVE@ < cpu_= core@CYCLE_ACTIVITY.STALLS_TOTAL@ - cpu_core@EXE_ACTIVITY.BOUND_ON_LOADS@ e= lse (cpu_core@EXE_ACTIVITY.1_PORTS_UTIL@ + tma_retiring * cpu_core@EXE_ACTI= VITY.2_3_PORTS_UTIL@) / tma_info_thread_clks)", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", @@ -2580,6 +2586,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "max(cpu_core@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ - cpu_= core@RESOURCE_STALLS.SCOREBOARD@, 0) / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", @@ -2590,6 +2597,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "cpu_core@EXE_ACTIVITY.1_PORTS_UTIL@ / tma_info_thre= ad_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", @@ -2600,7 +2608,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@EXE_ACTIVITY.2_PORTS_UTIL@ / tma_info_thre= ad_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", @@ -2611,7 +2618,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 3 or more uops per cycle on all execution ports (Logical= Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@UOPS_EXECUTED.CYCLES_GE_3@ / tma_info_thre= ad_clks", "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_ut= ilization_group", "MetricName": "tma_ports_utilized_3m", @@ -2632,7 +2638,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "cpu_core@topdown\\-retiring@ / (cpu_core@topdown\\-= fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@= + cpu_core@topdown\\-be\\-bound@) + 0 * tma_info_thread_slots", + "MetricExpr": "cpu_core@topdown\\-retiring@ / (cpu_core@topdown\\-= fe\\-bound@ + cpu_core@topdown\\-bad\\-spec@ + cpu_core@topdown\\-retiring@= + cpu_core@topdown\\-be\\-bound@)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -2663,7 +2669,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to PAUSE Instructions", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "cpu_core@CPU_CLK_UNHALTED.PAUSE@ / tma_info_thread_= clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_slow_pause", @@ -2698,7 +2703,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%", "Unit": "cpu_core" }, diff --git a/tools/perf/pmu-events/arch/x86/meteorlake/other.json b/tools/p= erf/pmu-events/arch/x86/meteorlake/other.json index cb21bb933617..8320ffd83c51 100644 --- a/tools/perf/pmu-events/arch/x86/meteorlake/other.json +++ b/tools/perf/pmu-events/arch/x86/meteorlake/other.json @@ -4,7 +4,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.HARDWARE", - "PublicDescription": "Count all other hardware assists or traps th= at are not necessarily architecturally exposed (through a software handler)= beyond FP; SSE-AVX mix and A/D assists who are counted by dedicated sub-ev= ents. This includes, but not limited to, assists at EXE or MEM uop writeba= ck like AVX* load/store/gather/scatter (non-FP GSSE-assist ) , assists gene= rated by ROB like PEBS and RTIT, Uncore trap, RAR (Remote Action Request) a= nd CET (Control flow Enforcement Technology) assists. the event also counts= for Machine Ordering count. Available PDIST counters: 0", + "PublicDescription": "Count all other hardware assists or traps th= at are not necessarily architecturally exposed (through a software handler)= beyond FP; SSE-AVX mix and A/D assists who are counted by dedicated sub-ev= ents. This includes, but not limited to, assists at EXE or MEM uop writeba= ck like AVX* load/store/gather/scatter (non-FP GSSE-assist ) , assists gene= rated by ROB like PEBS and RTIT, Uncore trap, RAR (Remote Action Request) a= nd CET (Control flow Enforcement Technology) assists. the event also counts= for Machine Ordering count.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -14,7 +14,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.PAGE_FAULT", - "PublicDescription": "ASSISTS.PAGE_FAULT Available PDIST counters:= 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -83,7 +82,7 @@ "CounterMask": "1", "EventCode": "0x2d", "EventName": "XQ.FULL_CYCLES", - "PublicDescription": "number of cycles when the thread is active a= nd the uncore cannot take any further requests (for example prefetches, loa= ds or stores initiated by the Core that miss the L2 cache). Available PDIST= counters: 0", + "PublicDescription": "number of cycles when the thread is active a= nd the uncore cannot take any further requests (for example prefetches, loa= ds or stores initiated by the Core that miss the L2 cache).", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/meteorlake/pipeline.json b/tool= s/perf/pmu-events/arch/x86/meteorlake/pipeline.json index 22b25708e799..bfdaabe9377d 100644 --- a/tools/perf/pmu-events/arch/x86/meteorlake/pipeline.json +++ b/tools/perf/pmu-events/arch/x86/meteorlake/pipeline.json @@ -15,7 +15,7 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.DIV_ACTIVE", - "PublicDescription": "Counts cycles when divide unit is busy execu= ting divide or square root operations. Accounts for integer and floating-po= int operations. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when divide unit is busy execu= ting divide or square root operations. Accounts for integer and floating-po= int operations.", "SampleAfterValue": "1000003", "UMask": "0x9", "Unit": "cpu_core" @@ -26,7 +26,6 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.IDIV_ACTIVE", - "PublicDescription": "This event counts the cycles the integer div= ider is busy. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -36,7 +35,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.ANY", - "PublicDescription": "Counts the number of occurrences where a mic= rocode assist is invoked by hardware. Examples include AD (page Access Dirt= y), FP and AVX related assists. Available PDIST counters: 0", + "PublicDescription": "Counts the number of occurrences where a mic= rocode assist is invoked by hardware. Examples include AD (page Access Dirt= y), FP and AVX related assists.", "SampleAfterValue": "100003", "UMask": "0x1b", "Unit": "cpu_core" @@ -44,6 +43,7 @@ { "BriefDescription": "Counts the total number of branch instruction= s retired for all branch types.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "MTL012, MTL013", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.ALL_BRANCHES", "PublicDescription": "Counts the total number of instructions in w= hich the instruction pointer (IP) of the processor is resteered due to a br= anch instruction and the branch instruction successfully retires. All bran= ch type instructions are accounted for.", @@ -62,6 +62,7 @@ { "BriefDescription": "Counts the number of retired JCC (Jump on Con= ditional Code) branch instructions retired, includes both taken and not tak= en branches.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "MTL013", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.COND", "SampleAfterValue": "200003", @@ -110,6 +111,7 @@ { "BriefDescription": "Counts the number of far branch instructions = retired, includes far jump, far call and return, and interrupt call and ret= urn.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "MTL013", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.FAR_BRANCH", "SampleAfterValue": "200003", @@ -129,6 +131,7 @@ { "BriefDescription": "Counts the number of near indirect JMP and ne= ar indirect CALL branch instructions retired.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "MTL013", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.INDIRECT", "SampleAfterValue": "200003", @@ -148,6 +151,7 @@ { "BriefDescription": "Counts the number of near indirect CALL branc= h instructions retired.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "MTL013", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.INDIRECT_CALL", "SampleAfterValue": "200003", @@ -167,6 +171,7 @@ "BriefDescription": "This event is deprecated. Refer to new event = BR_INST_RETIRED.INDIRECT_CALL", "Counter": "0,1,2,3,4,5,6,7", "Deprecated": "1", + "Errata": "MTL013", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.IND_CALL", "SampleAfterValue": "200003", @@ -176,6 +181,7 @@ { "BriefDescription": "Counts the number of near CALL branch instruc= tions retired.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "MTL012, MTL013", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_CALL", "SampleAfterValue": "200003", @@ -214,6 +220,7 @@ { "BriefDescription": "Counts the number of near taken branch instru= ctions retired.", "Counter": "0,1,2,3,4,5,6,7", + "Errata": "MTL013", "EventCode": "0xc4", "EventName": "BR_INST_RETIRED.NEAR_TAKEN", "SampleAfterValue": "200003", @@ -484,7 +491,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C01", - "PublicDescription": "Counts core clocks when the thread is in the= C0.1 light-weight slower wakeup time but more power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions. Availab= le PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.1 light-weight slower wakeup time but more power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions.", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" @@ -494,7 +501,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C02", - "PublicDescription": "Counts core clocks when the thread is in the= C0.2 light-weight faster wakeup time but less power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions. Availab= le PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.2 light-weight faster wakeup time but less power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions.", "SampleAfterValue": "2000003", "UMask": "0x20", "Unit": "cpu_core" @@ -504,7 +511,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C0_WAIT", - "PublicDescription": "Counts core clocks when the thread is in the= C0.1 or C0.2 power saving optimized states (TPAUSE or UMWAIT instructions)= or running the PAUSE instruction. Available PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.1 or C0.2 power saving optimized states (TPAUSE or UMWAIT instructions)= or running the PAUSE instruction.", "SampleAfterValue": "2000003", "UMask": "0x70", "Unit": "cpu_core" @@ -530,7 +537,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.DISTRIBUTED", - "PublicDescription": "This event distributes cycle counts between = active hyperthreads, i.e., those in C0. A hyperthread becomes inactive whe= n it executes the HLT or MWAIT instructions. If all other hyperthreads are= inactive (or disabled or do not exist), all counts are attributed to this = hyperthread. To obtain the full count when the Core is active, sum the coun= ts from each hyperthread. Available PDIST counters: 0", + "PublicDescription": "This event distributes cycle counts between = active hyperthreads, i.e., those in C0. A hyperthread becomes inactive whe= n it executes the HLT or MWAIT instructions. If all other hyperthreads are= inactive (or disabled or do not exist), all counts are attributed to this = hyperthread. To obtain the full count when the Core is active, sum the coun= ts from each hyperthread.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -540,7 +547,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE", - "PublicDescription": "Counts Core crystal clock cycles when curren= t thread is unhalted and the other thread is halted. Available PDIST counte= rs: 0", + "PublicDescription": "Counts Core crystal clock cycles when curren= t thread is unhalted and the other thread is halted.", "SampleAfterValue": "25003", "UMask": "0x2", "Unit": "cpu_core" @@ -550,7 +557,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.PAUSE", - "PublicDescription": "CPU_CLK_UNHALTED.PAUSE Available PDIST count= ers: 0", "SampleAfterValue": "2000003", "UMask": "0x40", "Unit": "cpu_core" @@ -562,7 +568,6 @@ "EdgeDetect": "1", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.PAUSE_INST", - "PublicDescription": "CPU_CLK_UNHALTED.PAUSE_INST Available PDIST = counters: 0", "SampleAfterValue": "2000003", "UMask": "0x40", "Unit": "cpu_core" @@ -572,7 +577,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.REF_DISTRIBUTED", - "PublicDescription": "This event distributes Core crystal clock cy= cle counts between active hyperthreads, i.e., those in C0 sleep-state. A hy= perthread becomes inactive when it executes the HLT or MWAIT instructions. = If one thread is active in a core, all counts are attributed to this hypert= hread. To obtain the full count when the Core is active, sum the counts fro= m each hyperthread. Available PDIST counters: 0", + "PublicDescription": "This event distributes Core crystal clock cy= cle counts between active hyperthreads, i.e., those in C0 sleep-state. A hy= perthread becomes inactive when it executes the HLT or MWAIT instructions. = If one thread is active in a core, all counts are attributed to this hypert= hread. To obtain the full count when the Core is active, sum the counts fro= m each hyperthread.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -589,7 +594,7 @@ "BriefDescription": "Reference cycles when the core is not in halt= state.", "Counter": "Fixed counter 2", "EventName": "CPU_CLK_UNHALTED.REF_TSC", - "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the eight programmable counte= rs available for other events. Note: On all current platforms this event st= ops counting during 'throttling (TM)' states duty off periods the processor= is 'halted'. The counter update is done at a lower clock rate then the co= re clock the overflow status bit for this counter may appear 'sticky'. Aft= er the counter has overflowed and software clears the overflow status bit a= nd resets the counter to less than MAX. The reset value to the counter is n= ot clocked immediately so the overflow status bit will flip 'high (1)' and = generate another PMI (if enabled) after which the reset value gets clocked = into the counter. Therefore, software will get the interrupt, read the over= flow status bit '1 for bit 34 while the counter value is less than MAX. Sof= tware should ignore this case. Available PDIST counters: 0", + "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the eight programmable counte= rs available for other events. Note: On all current platforms this event st= ops counting during 'throttling (TM)' states duty off periods the processor= is 'halted'. The counter update is done at a lower clock rate then the co= re clock the overflow status bit for this counter may appear 'sticky'. Aft= er the counter has overflowed and software clears the overflow status bit a= nd resets the counter to less than MAX. The reset value to the counter is n= ot clocked immediately so the overflow status bit will flip 'high (1)' and = generate another PMI (if enabled) after which the reset value gets clocked = into the counter. Therefore, software will get the interrupt, read the over= flow status bit '1 for bit 34 while the counter value is less than MAX. Sof= tware should ignore this case.", "SampleAfterValue": "2000003", "UMask": "0x3", "Unit": "cpu_core" @@ -609,7 +614,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.REF_TSC_P", - "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the four (eight when Hyperthr= eading is disabled) programmable counters available for other events. Note:= On all current platforms this event stops counting during 'throttling (TM)= ' states duty off periods the processor is 'halted'. The counter update is= done at a lower clock rate then the core clock the overflow status bit for= this counter may appear 'sticky'. After the counter has overflowed and so= ftware clears the overflow status bit and resets the counter to less than M= AX. The reset value to the counter is not clocked immediately so the overfl= ow status bit will flip 'high (1)' and generate another PMI (if enabled) af= ter which the reset value gets clocked into the counter. Therefore, softwar= e will get the interrupt, read the overflow status bit '1 for bit 34 while = the counter value is less than MAX. Software should ignore this case. Avail= able PDIST counters: 0", + "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the four (eight when Hyperthr= eading is disabled) programmable counters available for other events. Note:= On all current platforms this event stops counting during 'throttling (TM)= ' states duty off periods the processor is 'halted'. The counter update is= done at a lower clock rate then the core clock the overflow status bit for= this counter may appear 'sticky'. After the counter has overflowed and so= ftware clears the overflow status bit and resets the counter to less than M= AX. The reset value to the counter is not clocked immediately so the overfl= ow status bit will flip 'high (1)' and generate another PMI (if enabled) af= ter which the reset value gets clocked into the counter. Therefore, softwar= e will get the interrupt, read the overflow status bit '1 for bit 34 while = the counter value is less than MAX. Software should ignore this case.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -626,7 +631,7 @@ "BriefDescription": "Core cycles when the thread is not in halt st= ate", "Counter": "Fixed counter 1", "EventName": "CPU_CLK_UNHALTED.THREAD", - "PublicDescription": "Counts the number of core cycles while the t= hread is not in a halt state. The thread enters the halt state when it is r= unning the HLT instruction. This event is a component in many key event rat= ios. The core frequency may change from time to time due to transitions ass= ociated with Enhanced Intel SpeedStep Technology or TM2. For this reason th= is event may have a changing ratio with regards to time. When the core freq= uency is constant, this event can approximate elapsed time while the core w= as not in the halt state. It is counted on a dedicated fixed counter, leavi= ng the eight programmable counters available for other events. Available PD= IST counters: 0", + "PublicDescription": "Counts the number of core cycles while the t= hread is not in a halt state. The thread enters the halt state when it is r= unning the HLT instruction. This event is a component in many key event rat= ios. The core frequency may change from time to time due to transitions ass= ociated with Enhanced Intel SpeedStep Technology or TM2. For this reason th= is event may have a changing ratio with regards to time. When the core freq= uency is constant, this event can approximate elapsed time while the core w= as not in the halt state. It is counted on a dedicated fixed counter, leavi= ng the eight programmable counters available for other events.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -644,7 +649,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.THREAD_P", - "PublicDescription": "This is an architectural event that counts t= he number of thread cycles while the thread is not in a halt state. The thr= ead enters the halt state when it is running the HLT instruction. The core = frequency may change from time to time due to power or thermal throttling. = For this reason, this event may have a changing ratio with regards to wall = clock time. Available PDIST counters: 0", + "PublicDescription": "This is an architectural event that counts t= he number of thread cycles while the thread is not in a halt state. The thr= ead enters the halt state when it is running the HLT instruction. The core = frequency may change from time to time due to power or thermal throttling. = For this reason, this event may have a changing ratio with regards to wall = clock time.", "SampleAfterValue": "2000003", "Unit": "cpu_core" }, @@ -654,7 +659,6 @@ "CounterMask": "8", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS", - "PublicDescription": "Cycles while L1 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8", "Unit": "cpu_core" @@ -665,7 +669,6 @@ "CounterMask": "1", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS", - "PublicDescription": "Cycles while L2 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -676,7 +679,6 @@ "CounterMask": "16", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY", - "PublicDescription": "Cycles while memory subsystem has an outstan= ding load. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -687,7 +689,6 @@ "CounterMask": "12", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS", - "PublicDescription": "Execution stalls while L1 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0xc", "Unit": "cpu_core" @@ -698,7 +699,6 @@ "CounterMask": "5", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS", - "PublicDescription": "Execution stalls while L2 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x5", "Unit": "cpu_core" @@ -709,7 +709,6 @@ "CounterMask": "4", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL", - "PublicDescription": "Total execution stalls. Available PDIST coun= ters: 0", "SampleAfterValue": "1000003", "UMask": "0x4", "Unit": "cpu_core" @@ -719,7 +718,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.1_PORTS_UTIL", - "PublicDescription": "Counts cycles during which a total of 1 uop = was executed on all ports and Reservation Station (RS) was not empty. Avail= able PDIST counters: 0", + "PublicDescription": "Counts cycles during which a total of 1 uop = was executed on all ports and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -729,7 +728,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.2_3_PORTS_UTIL", - "PublicDescription": "Cycles total of 2 or 3 uops are executed on = all ports and Reservation Station (RS) was not empty. Available PDIST count= ers: 0", "SampleAfterValue": "2000003", "UMask": "0xc", "Unit": "cpu_core" @@ -739,7 +737,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.2_PORTS_UTIL", - "PublicDescription": "Counts cycles during which a total of 2 uops= were executed on all ports and Reservation Station (RS) was not empty. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts cycles during which a total of 2 uops= were executed on all ports and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -749,7 +747,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.3_PORTS_UTIL", - "PublicDescription": "Cycles total of 3 uops are executed on all p= orts and Reservation Station (RS) was not empty. Available PDIST counters: = 0", + "PublicDescription": "Cycles total of 3 uops are executed on all p= orts and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -759,7 +757,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.4_PORTS_UTIL", - "PublicDescription": "Cycles total of 4 uops are executed on all p= orts and Reservation Station (RS) was not empty. Available PDIST counters: = 0", + "PublicDescription": "Cycles total of 4 uops are executed on all p= orts and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" @@ -770,7 +768,6 @@ "CounterMask": "5", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.BOUND_ON_LOADS", - "PublicDescription": "Execution stalls while memory subsystem has = an outstanding load. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x21", "Unit": "cpu_core" @@ -781,7 +778,7 @@ "CounterMask": "2", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.BOUND_ON_STORES", - "PublicDescription": "Counts cycles where the Store Buffer was ful= l and no loads caused an execution stall. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where the Store Buffer was ful= l and no loads caused an execution stall.", "SampleAfterValue": "1000003", "UMask": "0x40", "Unit": "cpu_core" @@ -791,7 +788,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.EXE_BOUND_0_PORTS", - "PublicDescription": "Number of cycles total of 0 uops executed on= all ports, Reservation Station (RS) was not empty, the Store Buffer (SB) w= as not full and there was no outstanding load. Available PDIST counters: 0", + "PublicDescription": "Number of cycles total of 0 uops executed on= all ports, Reservation Station (RS) was not empty, the Store Buffer (SB) w= as not full and there was no outstanding load.", "SampleAfterValue": "1000003", "UMask": "0x80", "Unit": "cpu_core" @@ -801,7 +798,7 @@ "Counter": "0,1,2,3", "EventCode": "0x75", "EventName": "INST_DECODED.DECODERS", - "PublicDescription": "Number of decoders utilized in a cycle when = the MITE (legacy decode pipeline) fetches instructions. Available PDIST cou= nters: 0", + "PublicDescription": "Number of decoders utilized in a cycle when = the MITE (legacy decode pipeline) fetches instructions.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -846,7 +843,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.MACRO_FUSED", - "PublicDescription": "INST_RETIRED.MACRO_FUSED Available PDIST cou= nters: 0", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" @@ -856,7 +852,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.NOP", - "PublicDescription": "Counts all retired NOP or ENDBR32/64 or PREF= ETCHIT0/1 instructions Available PDIST counters: 0", + "PublicDescription": "Counts all retired NOP or ENDBR32/64 or PREF= ETCHIT0/1 instructions", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -875,7 +871,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.REP_ITERATION", - "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent. Available PDIST counters: 0", + "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent.", "SampleAfterValue": "2000003", "UMask": "0x8", "Unit": "cpu_core" @@ -887,7 +883,7 @@ "EdgeDetect": "1", "EventCode": "0xad", "EventName": "INT_MISC.CLEARS_COUNT", - "PublicDescription": "Counts the number of speculative clears due = to any type of branch misprediction or machine clears Available PDIST count= ers: 0", + "PublicDescription": "Counts the number of speculative clears due = to any type of branch misprediction or machine clears", "SampleAfterValue": "500009", "UMask": "0x1", "Unit": "cpu_core" @@ -897,7 +893,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.CLEAR_RESTEER_CYCLES", - "PublicDescription": "Cycles after recovery from a branch mispredi= ction or machine clear till the first uop is issued from the resteered path= . Available PDIST counters: 0", + "PublicDescription": "Cycles after recovery from a branch mispredi= ction or machine clear till the first uop is issued from the resteered path= .", "SampleAfterValue": "500009", "UMask": "0x80", "Unit": "cpu_core" @@ -907,7 +903,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.RECOVERY_CYCLES", - "PublicDescription": "Counts core cycles when the Resource allocat= or was stalled due to recovery from an earlier branch misprediction or mach= ine clear event. Available PDIST counters: 0", + "PublicDescription": "Counts core cycles when the Resource allocat= or was stalled due to recovery from an earlier branch misprediction or mach= ine clear event.", "SampleAfterValue": "500009", "UMask": "0x1", "Unit": "cpu_core" @@ -919,7 +915,6 @@ "EventName": "INT_MISC.UNKNOWN_BRANCH_CYCLES", "MSRIndex": "0x3F7", "MSRValue": "0x7", - "PublicDescription": "Bubble cycles of BAClear (Unknown Branch). A= vailable PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x40", "Unit": "cpu_core" @@ -929,7 +924,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.UOP_DROPPING", - "PublicDescription": "Estimated number of Top-down Microarchitectu= re Analysis slots that got dropped due to non front-end reasons Available P= DIST counters: 0", + "PublicDescription": "Estimated number of Top-down Microarchitectu= re Analysis slots that got dropped due to non front-end reasons", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -939,7 +934,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.128BIT", - "PublicDescription": "INT_VEC_RETIRED.128BIT Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0x13", "Unit": "cpu_core" @@ -949,7 +943,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.256BIT", - "PublicDescription": "INT_VEC_RETIRED.256BIT Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0xac", "Unit": "cpu_core" @@ -959,7 +952,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.ADD_128", - "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 128-bit vector instructions. Available PDIST counters: 0= ", + "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 128-bit vector instructions.", "SampleAfterValue": "1000003", "UMask": "0x3", "Unit": "cpu_core" @@ -969,7 +962,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.ADD_256", - "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 256-bit vector instructions. Available PDIST counters: 0= ", + "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 256-bit vector instructions.", "SampleAfterValue": "1000003", "UMask": "0xc", "Unit": "cpu_core" @@ -979,7 +972,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.MUL_256", - "PublicDescription": "INT_VEC_RETIRED.MUL_256 Available PDIST coun= ters: 0", "SampleAfterValue": "1000003", "UMask": "0x80", "Unit": "cpu_core" @@ -989,7 +981,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.SHUFFLES", - "PublicDescription": "INT_VEC_RETIRED.SHUFFLES Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x40", "Unit": "cpu_core" @@ -999,7 +990,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.VNNI_128", - "PublicDescription": "INT_VEC_RETIRED.VNNI_128 Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x10", "Unit": "cpu_core" @@ -1009,7 +999,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.VNNI_256", - "PublicDescription": "INT_VEC_RETIRED.VNNI_256 Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x20", "Unit": "cpu_core" @@ -1028,7 +1017,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.ADDRESS_ALIAS", - "PublicDescription": "Counts the number of times a load got blocke= d due to false dependencies in MOB due to partial compare on address. Avail= able PDIST counters: 0", + "PublicDescription": "Counts the number of times a load got blocke= d due to false dependencies in MOB due to partial compare on address.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -1047,7 +1036,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.NO_SR", - "PublicDescription": "Counts the number of times that split load o= perations are temporarily blocked because all resources for handling the sp= lit accesses are in use. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times that split load o= perations are temporarily blocked because all resources for handling the sp= lit accesses are in use.", "SampleAfterValue": "100003", "UMask": "0x88", "Unit": "cpu_core" @@ -1066,7 +1055,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.STORE_FORWARD", - "PublicDescription": "Counts the number of times where store forwa= rding was prevented for a load operation. The most common case is a load bl= ocked due to the address of memory access (partially) overlapping with a pr= eceding uncompleted store. Note: See the table of not supported store forwa= rds in the Optimization Guide. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times where store forwa= rding was prevented for a load operation. The most common case is a load bl= ocked due to the address of memory access (partially) overlapping with a pr= eceding uncompleted store. Note: See the table of not supported store forwa= rds in the Optimization Guide.", "SampleAfterValue": "100003", "UMask": "0x82", "Unit": "cpu_core" @@ -1076,7 +1065,7 @@ "Counter": "0,1,2,3", "EventCode": "0x4c", "EventName": "LOAD_HIT_PREFETCH.SWPF", - "PublicDescription": "Counts all software-prefetch load dispatches= that hit the fill buffer (FB) allocated for the software prefetch. It can = also be incremented by some lock instructions. So it should only be used wi= th profiling so that the locks can be excluded by ASM (Assembly File) inspe= ction of the nearby instructions. Available PDIST counters: 0", + "PublicDescription": "Counts all software-prefetch load dispatches= that hit the fill buffer (FB) allocated for the software prefetch. It can = also be incremented by some lock instructions. So it should only be used wi= th profiling so that the locks can be excluded by ASM (Assembly File) inspe= ction of the nearby instructions.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -1087,7 +1076,7 @@ "CounterMask": "1", "EventCode": "0xa8", "EventName": "LSD.CYCLES_ACTIVE", - "PublicDescription": "Counts the cycles when at least one uop is d= elivered by the LSD (Loop-stream detector). Available PDIST counters: 0", + "PublicDescription": "Counts the cycles when at least one uop is d= elivered by the LSD (Loop-stream detector).", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1098,7 +1087,7 @@ "CounterMask": "6", "EventCode": "0xa8", "EventName": "LSD.CYCLES_OK", - "PublicDescription": "Counts the cycles when optimal number of uop= s is delivered by the LSD (Loop-stream detector). Available PDIST counters:= 0", + "PublicDescription": "Counts the cycles when optimal number of uop= s is delivered by the LSD (Loop-stream detector).", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1108,7 +1097,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa8", "EventName": "LSD.UOPS", - "PublicDescription": "Counts the number of uops delivered to the b= ack-end by the LSD(Loop Stream Detector). Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops delivered to the b= ack-end by the LSD(Loop Stream Detector).", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1128,7 +1117,7 @@ "EdgeDetect": "1", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.COUNT", - "PublicDescription": "Counts the number of machine clears (nukes) = of any type. Available PDIST counters: 0", + "PublicDescription": "Counts the number of machine clears (nukes) = of any type.", "SampleAfterValue": "100003", "UMask": "0x1", "Unit": "cpu_core" @@ -1184,7 +1173,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.SMC", - "PublicDescription": "Counts self-modifying code (SMC) detected, w= hich causes a machine clear. Available PDIST counters: 0", + "PublicDescription": "Counts self-modifying code (SMC) detected, w= hich causes a machine clear.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -1194,7 +1183,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe0", "EventName": "MISC2_RETIRED.LFENCE", - "PublicDescription": "number of LFENCE retired instructions Availa= ble PDIST counters: 0", + "PublicDescription": "number of LFENCE retired instructions", "SampleAfterValue": "400009", "UMask": "0x20", "Unit": "cpu_core" @@ -1213,7 +1202,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcc", "EventName": "MISC_RETIRED.LBR_INSERTS", - "PublicDescription": "Increments when an entry is added to the Las= t Branch Record (LBR) array (or removed from the array in case of RETURNs i= n call stack mode). The event requires LBR enable via IA32_DEBUGCTL MSR and= branch type selection via MSR_LBR_SELECT. Available PDIST counters: 0", + "PublicDescription": "Increments when an entry is added to the Las= t Branch Record (LBR) array (or removed from the array in case of RETURNs i= n call stack mode). The event requires LBR enable via IA32_DEBUGCTL MSR and= branch type selection via MSR_LBR_SELECT.", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -1223,7 +1212,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa2", "EventName": "RESOURCE_STALLS.SB", - "PublicDescription": "Counts allocation stall cycles caused by the= store buffer (SB) being full. This counts cycles that the pipeline back-en= d blocked uop delivery from the front-end. Available PDIST counters: 0", + "PublicDescription": "Counts allocation stall cycles caused by the= store buffer (SB) being full. This counts cycles that the pipeline back-en= d blocked uop delivery from the front-end.", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -1233,7 +1222,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa2", "EventName": "RESOURCE_STALLS.SCOREBOARD", - "PublicDescription": "Counts cycles where the pipeline is stalled = due to serializing operations. Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -1243,7 +1231,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa5", "EventName": "RS.EMPTY", - "PublicDescription": "Counts cycles during which the reservation s= tation (RS) is empty for this logical processor. This is usually caused whe= n the front-end pipeline runs into starvation periods (e.g. branch mispredi= ctions or i-cache misses) Available PDIST counters: 0", + "PublicDescription": "Counts cycles during which the reservation s= tation (RS) is empty for this logical processor. This is usually caused whe= n the front-end pipeline runs into starvation periods (e.g. branch mispredi= ctions or i-cache misses)", "SampleAfterValue": "1000003", "UMask": "0x7", "Unit": "cpu_core" @@ -1256,7 +1244,7 @@ "EventCode": "0xa5", "EventName": "RS.EMPTY_COUNT", "Invert": "1", - "PublicDescription": "Counts end of periods where the Reservation = Station (RS) was empty. Could be useful to closely sample on front-end late= ncy issues (see the FRONTEND_RETIRED event of designated precise events) Av= ailable PDIST counters: 0", + "PublicDescription": "Counts end of periods where the Reservation = Station (RS) was empty. Could be useful to closely sample on front-end late= ncy issues (see the FRONTEND_RETIRED event of designated precise events)", "SampleAfterValue": "100003", "UMask": "0x7", "Unit": "cpu_core" @@ -1266,7 +1254,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa5", "EventName": "RS.EMPTY_RESOURCE", - "PublicDescription": "Cycles when RS was empty and a resource allo= cation stall is asserted Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1285,7 +1272,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.BACKEND_BOUND_SLOTS", - "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that were not consumed by the back-end pipeline due to lack of ba= ck-end resources, as a result of memory subsystem delays, execution units l= imitations, or other conditions. The count is distributed among unhalted lo= gical processors (hyper-threads) who share the same physical core, in proce= ssors that support Intel Hyper-Threading Technology. Software can use this = event as the numerator for the Backend Bound metric (or top-level category)= of the Top-down Microarchitecture Analysis method. Available PDIST counter= s: 0", + "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that were not consumed by the back-end pipeline due to lack of ba= ck-end resources, as a result of memory subsystem delays, execution units l= imitations, or other conditions. The count is distributed among unhalted lo= gical processors (hyper-threads) who share the same physical core, in proce= ssors that support Intel Hyper-Threading Technology. Software can use this = event as the numerator for the Backend Bound metric (or top-level category)= of the Top-down Microarchitecture Analysis method.", "SampleAfterValue": "10000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1295,7 +1282,7 @@ "Counter": "0", "EventCode": "0xa4", "EventName": "TOPDOWN.BAD_SPEC_SLOTS", - "PublicDescription": "Number of slots of TMA method that were wast= ed due to incorrect speculation. It covers all types of control-flow or dat= a-related mis-speculations. Available PDIST counters: 0", + "PublicDescription": "Number of slots of TMA method that were wast= ed due to incorrect speculation. It covers all types of control-flow or dat= a-related mis-speculations.", "SampleAfterValue": "10000003", "UMask": "0x4", "Unit": "cpu_core" @@ -1305,7 +1292,7 @@ "Counter": "0", "EventCode": "0xa4", "EventName": "TOPDOWN.BR_MISPREDICT_SLOTS", - "PublicDescription": "Number of TMA slots that were wasted due to = incorrect speculation by (any type of) branch mispredictions. This event es= timates number of speculative operations that were issued but not retired a= s well as the out-of-order engine recovery past a branch misprediction. Ava= ilable PDIST counters: 0", + "PublicDescription": "Number of TMA slots that were wasted due to = incorrect speculation by (any type of) branch mispredictions. This event es= timates number of speculative operations that were issued but not retired a= s well as the out-of-order engine recovery past a branch misprediction.", "SampleAfterValue": "10000003", "UMask": "0x8", "Unit": "cpu_core" @@ -1315,7 +1302,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.MEMORY_BOUND_SLOTS", - "PublicDescription": "TOPDOWN.MEMORY_BOUND_SLOTS Available PDIST c= ounters: 0", "SampleAfterValue": "10000003", "UMask": "0x10", "Unit": "cpu_core" @@ -1324,7 +1310,7 @@ "BriefDescription": "TMA slots available for an unhalted logical p= rocessor. Fixed counter - architectural event", "Counter": "Fixed counter 3", "EventName": "TOPDOWN.SLOTS", - "PublicDescription": "Number of available slots for an unhalted lo= gical processor. The event increments by machine-width of the narrowest pip= eline as employed by the Top-down Microarchitecture Analysis method (TMA). = The count is distributed among unhalted logical processors (hyper-threads) = who share the same physical core. Software can use this event as the denomi= nator for the top-level metrics of the TMA method. This architectural event= is counted on a designated fixed counter (Fixed Counter 3). Available PDIS= T counters: 0", + "PublicDescription": "Number of available slots for an unhalted lo= gical processor. The event increments by machine-width of the narrowest pip= eline as employed by the Top-down Microarchitecture Analysis method (TMA). = The count is distributed among unhalted logical processors (hyper-threads) = who share the same physical core. Software can use this event as the denomi= nator for the top-level metrics of the TMA method. This architectural event= is counted on a designated fixed counter (Fixed Counter 3).", "SampleAfterValue": "10000003", "UMask": "0x4", "Unit": "cpu_core" @@ -1334,7 +1320,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.SLOTS_P", - "PublicDescription": "Counts the number of available slots for an = unhalted logical processor. The event increments by machine-width of the na= rrowest pipeline as employed by the Top-down Microarchitecture Analysis met= hod. The count is distributed among unhalted logical processors (hyper-thre= ads) who share the same physical core. Available PDIST counters: 0", + "PublicDescription": "Counts the number of available slots for an = unhalted logical processor. The event increments by machine-width of the na= rrowest pipeline as employed by the Top-down Microarchitecture Analysis met= hod. The count is distributed among unhalted logical processors (hyper-thre= ads) who share the same physical core.", "SampleAfterValue": "10000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1591,7 +1577,7 @@ "Counter": "0,1,2,3", "EventCode": "0x76", "EventName": "UOPS_DECODED.DEC0_UOPS", - "PublicDescription": "This event counts the number of not dec-by-a= ll uops decoded by decoder 0. Available PDIST counters: 0", + "PublicDescription": "This event counts the number of not dec-by-a= ll uops decoded by decoder 0.", "SampleAfterValue": "1000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1601,7 +1587,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_0", - "PublicDescription": "Number of uops dispatch to execution port 0= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 0= .", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1611,7 +1597,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_1", - "PublicDescription": "Number of uops dispatch to execution port 1= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 1= .", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1621,7 +1607,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_2_3_10", - "PublicDescription": "Number of uops dispatch to execution ports 2= , 3 and 10 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 2= , 3 and 10", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -1631,7 +1617,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_4_9", - "PublicDescription": "Number of uops dispatch to execution ports 4= and 9 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 4= and 9", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" @@ -1641,7 +1627,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_5_11", - "PublicDescription": "Number of uops dispatch to execution ports 5= and 11 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 5= and 11", "SampleAfterValue": "2000003", "UMask": "0x20", "Unit": "cpu_core" @@ -1651,7 +1637,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_6", - "PublicDescription": "Number of uops dispatch to execution port 6= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 6= .", "SampleAfterValue": "2000003", "UMask": "0x40", "Unit": "cpu_core" @@ -1661,7 +1647,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_7_8", - "PublicDescription": "Number of uops dispatch to execution ports = 7 and 8. Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports = 7 and 8.", "SampleAfterValue": "2000003", "UMask": "0x80", "Unit": "cpu_core" @@ -1671,7 +1657,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE", - "PublicDescription": "Counts the number of uops executed from any = thread. Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops executed from any = thread.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1682,7 +1668,7 @@ "CounterMask": "1", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1", - "PublicDescription": "Counts cycles when at least 1 micro-op is ex= ecuted from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 1 micro-op is ex= ecuted from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1693,7 +1679,7 @@ "CounterMask": "2", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2", - "PublicDescription": "Counts cycles when at least 2 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 2 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1704,7 +1690,7 @@ "CounterMask": "3", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3", - "PublicDescription": "Counts cycles when at least 3 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 3 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1715,7 +1701,7 @@ "CounterMask": "4", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4", - "PublicDescription": "Counts cycles when at least 4 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 4 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1726,7 +1712,7 @@ "CounterMask": "1", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_1", - "PublicDescription": "Cycles where at least 1 uop was executed per= -thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 1 uop was executed per= -thread.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1737,7 +1723,7 @@ "CounterMask": "2", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_2", - "PublicDescription": "Cycles where at least 2 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 2 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1748,7 +1734,7 @@ "CounterMask": "3", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_3", - "PublicDescription": "Cycles where at least 3 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 3 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1759,7 +1745,7 @@ "CounterMask": "4", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_4", - "PublicDescription": "Cycles where at least 4 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 4 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1771,7 +1757,7 @@ "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.STALLS", "Invert": "1", - "PublicDescription": "Counts cycles during which no uops were disp= atched from the Reservation Station (RS) per thread. Available PDIST counte= rs: 0", + "PublicDescription": "Counts cycles during which no uops were disp= atched from the Reservation Station (RS) per thread.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1781,7 +1767,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.THREAD", - "PublicDescription": "Counts the number of uops to be executed per= -thread each cycle. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1791,7 +1776,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.X87", - "PublicDescription": "Counts the number of x87 uops executed. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts the number of x87 uops executed.", "SampleAfterValue": "2000003", "UMask": "0x10", "Unit": "cpu_core" @@ -1810,7 +1795,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xae", "EventName": "UOPS_ISSUED.ANY", - "PublicDescription": "Counts the number of uops that the Resource = Allocation Table (RAT) issues to the Reservation Station (RS). Available PD= IST counters: 0", + "PublicDescription": "Counts the number of uops that the Resource = Allocation Table (RAT) issues to the Reservation Station (RS).", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1821,7 +1806,6 @@ "CounterMask": "1", "EventCode": "0xae", "EventName": "UOPS_ISSUED.CYCLES", - "PublicDescription": "UOPS_ISSUED.CYCLES Available PDIST counters:= 0", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1840,7 +1824,7 @@ "CounterMask": "1", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.CYCLES", - "PublicDescription": "Counts cycles where at least one uop has ret= ired. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where at least one uop has ret= ired.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1850,7 +1834,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.HEAVY", - "PublicDescription": "Counts the number of retired micro-operation= s (uops) except the last uop of each instruction. An instruction that is de= coded into less than two uops does not contribute to the count. Available P= DIST counters: 0", + "PublicDescription": "Counts the number of retired micro-operation= s (uops) except the last uop of each instruction. An instruction that is de= coded into less than two uops does not contribute to the count.", "SampleAfterValue": "2000003", "UMask": "0x1", "Unit": "cpu_core" @@ -1880,7 +1864,6 @@ "EventName": "UOPS_RETIRED.MS", "MSRIndex": "0x3F7", "MSRValue": "0x8", - "PublicDescription": "UOPS_RETIRED.MS Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4", "Unit": "cpu_core" @@ -1890,7 +1873,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.SLOTS", - "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that are utilized by operations that eventually get retired (comm= itted) by the processor pipeline. Usually, this event positively correlates= with higher performance for example, as measured by the instructions-per-= cycle metric. Software can use this event as the numerator for the Retiring= metric (or top-level category) of the Top-down Microarchitecture Analysis = method. Available PDIST counters: 0", + "PublicDescription": "This event counts a subset of the Topdown Sl= ots event that are utilized by operations that eventually get retired (comm= itted) by the processor pipeline. Usually, this event positively correlates= with higher performance for example, as measured by the instructions-per-= cycle metric. Software can use this event as the numerator for the Retiring= metric (or top-level category) of the Top-down Microarchitecture Analysis = method.", "SampleAfterValue": "2000003", "UMask": "0x2", "Unit": "cpu_core" @@ -1902,7 +1885,7 @@ "EventCode": "0xc2", "EventName": "UOPS_RETIRED.STALLS", "Invert": "1", - "PublicDescription": "This event counts cycles without actually re= tired uops. Available PDIST counters: 0", + "PublicDescription": "This event counts cycles without actually re= tired uops.", "SampleAfterValue": "1000003", "UMask": "0x2", "Unit": "cpu_core" diff --git a/tools/perf/pmu-events/arch/x86/meteorlake/virtual-memory.json = b/tools/perf/pmu-events/arch/x86/meteorlake/virtual-memory.json index f300129e9e2d..305b96b26a4e 100644 --- a/tools/perf/pmu-events/arch/x86/meteorlake/virtual-memory.json +++ b/tools/perf/pmu-events/arch/x86/meteorlake/virtual-memory.json @@ -13,7 +13,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.STLB_HIT", - "PublicDescription": "Counts loads that miss the DTLB (Data TLB) a= nd hit the STLB (Second level TLB). Available PDIST counters: 0", + "PublicDescription": "Counts loads that miss the DTLB (Data TLB) a= nd hit the STLB (Second level TLB).", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -24,7 +24,7 @@ "CounterMask": "1", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a demand load. Available PDIST cou= nters: 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a demand load.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -43,7 +43,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data loads. This implies it missed in the DTLB and furth= er levels of TLB. The page walk can end with or without a fault. Available = PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data loads. This implies it missed in the DTLB and furth= er levels of TLB. The page walk can end with or without a fault.", "SampleAfterValue": "100003", "UMask": "0xe", "Unit": "cpu_core" @@ -53,7 +53,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_1G", - "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -73,7 +73,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data loads. This implies address translations missed in the= DTLB and further levels of TLB. The page walk can end with or without a fa= ult. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data loads. This implies address translations missed in the= DTLB and further levels of TLB. The page walk can end with or without a fa= ult.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -93,7 +93,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -113,7 +113,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for a demand load in the PMH (Page Miss Handler) each cycle. Available PDIS= T counters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for a demand load in the PMH (Page Miss Handler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -132,7 +132,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.STLB_HIT", - "PublicDescription": "Counts stores that miss the DTLB (Data TLB) = and hit the STLB (2nd Level TLB). Available PDIST counters: 0", + "PublicDescription": "Counts stores that miss the DTLB (Data TLB) = and hit the STLB (2nd Level TLB).", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -143,7 +143,7 @@ "CounterMask": "1", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a store. Available PDIST counters:= 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a store.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -162,7 +162,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data stores. This implies it missed in the DTLB and furt= her levels of TLB. The page walk can end with or without a fault. Available= PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data stores. This implies it missed in the DTLB and furt= her levels of TLB. The page walk can end with or without a fault.", "SampleAfterValue": "100003", "UMask": "0xe", "Unit": "cpu_core" @@ -172,7 +172,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_1G", - "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t.", "SampleAfterValue": "100003", "UMask": "0x8", "Unit": "cpu_core" @@ -192,7 +192,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data stores. This implies address translations missed in th= e DTLB and further levels of TLB. The page walk can end with or without a f= ault. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data stores. This implies address translations missed in th= e DTLB and further levels of TLB. The page walk can end with or without a f= ault.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -212,7 +212,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t.", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -232,7 +232,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for a store in the PMH (Page Miss Handler) each cycle. Available PDIST coun= ters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for a store in the PMH (Page Miss Handler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -260,7 +260,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.STLB_HIT", - "PublicDescription": "Counts instruction fetch requests that miss = the ITLB (Instruction TLB) and hit the STLB (Second-level TLB). Available P= DIST counters: 0", + "PublicDescription": "Counts instruction fetch requests that miss = the ITLB (Instruction TLB) and hit the STLB (Second-level TLB).", "SampleAfterValue": "100003", "UMask": "0x20", "Unit": "cpu_core" @@ -271,7 +271,7 @@ "CounterMask": "1", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a code (instruction fetch) request= . Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a code (instruction fetch) request= .", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" @@ -291,7 +291,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes)= caused by a code fetch. This implies it missed in the ITLB (Instruction TL= B) and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes)= caused by a code fetch. This implies it missed in the ITLB (Instruction TL= B) and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0xe", "Unit": "cpu_core" @@ -311,7 +311,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M page size= s) caused by a code fetch. This implies it missed in the ITLB (Instruction = TLB) and further levels of TLB. The page walk can end with or without a fau= lt. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M page size= s) caused by a code fetch. This implies it missed in the ITLB (Instruction = TLB) and further levels of TLB. The page walk can end with or without a fau= lt.", "SampleAfterValue": "100003", "UMask": "0x4", "Unit": "cpu_core" @@ -331,7 +331,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K page sizes) = caused by a code fetch. This implies it missed in the ITLB (Instruction TLB= ) and further levels of TLB. The page walk can end with or without a fault.= Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K page sizes) = caused by a code fetch. This implies it missed in the ITLB (Instruction TLB= ) and further levels of TLB. The page walk can end with or without a fault.= ", "SampleAfterValue": "100003", "UMask": "0x2", "Unit": "cpu_core" @@ -351,7 +351,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for an outstanding code (instruction fetch) request in the PMH (Page Miss H= andler) each cycle. Available PDIST counters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for an outstanding code (instruction fetch) request in the PMH (Page Miss H= andler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10", "Unit": "cpu_core" --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pl1-f202.google.com (mail-pl1-f202.google.com [209.85.214.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 230E232BF27 for ; Mon, 18 Aug 2025 19:06:11 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.202 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543977; cv=none; b=CZaqYfDEYdgzvzytnIrMrtRmpUbuuUJsrdFOJJSxDjMT7K5K1BK/1GMEKdxu2DAKUEbFeOtiQ6CfSI583gAl36S4CfhA7yT3nUyySNmHr0HWK9e2JX/NWBLXk2gZL7d9C5CMa2I02NuhsbQ9pIksbwB6JJCj0dTDVwDlRnwbOUU= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543977; c=relaxed/simple; bh=LDMsEsOPf0m/bxnH35Ptk/DeGrPTNgx/6p5y86XvHcM=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=oQ3KnDaeSC/Y/ljalbBQW1u/e4mLBYVdlt4Qhw1bbnO5d6Is7PdJQr6liu+mo3EPNwe3UMaqOXe/i9cK8sGeSkpeGM+ZhnzIH9j9dADhrIa/agyjWcw59uJvMxLrnonX82JvyT5ZtYh/slrtrk8XB3OZDipF/vBrkxBT/ynbXYw= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=17njsg5o; arc=none smtp.client-ip=209.85.214.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="17njsg5o" Received: by mail-pl1-f202.google.com with SMTP id d9443c01a7336-24458067fdeso52752685ad.1 for ; Mon, 18 Aug 2025 12:06:11 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543971; x=1756148771; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=x8WrBygYy6pzEoWgBnQh3kFLnYEsQbwsrR/YgnrkztA=; b=17njsg5oQPAygaQc/bi7R52BwzTpPQBKm+Lg+mOaXETfnL9gKcb9q9W4+lMKqgJk3z 0/yTZcx2qpcqIz4q1gK4uriqsXAiYC2WmQEnFBlD1Dx0Xax87QwtV53EPNf6Dbdj3ubD CmAfdx1o1LKoqODH+OTSYMFjAeWuvHkiC9CNieWnN7iVAoSEFO7VURUQEFXINHOqAjl0 errqzA3Fd02lSuZ4aouDa5cgK6cnebbPB3vr1NekcIKs4fFkFmHdGvSS/LgpvFNSlju9 7h3nWm1XLR60HYS+SWLOrPqafvZ0uVOARdxgJQ2tO5Q1xeWVcA2FqHXd0WNtbR8BtXzu gxfA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543971; x=1756148771; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=x8WrBygYy6pzEoWgBnQh3kFLnYEsQbwsrR/YgnrkztA=; b=OI87fwACsL5+5F3XGzU27pdsHIzAD6CEDDYqzsxWVG7a2XyAm5n1GHcKN3cnuG3k31 h9AiqfWbfGrUyTQ3e4ZsQk8iUcOsmmEjlzT0rvG9N3TvtVMqPAnIWq0yBaoYnIvKVXPO 2PQJM9KEUbCr0qGXosljwwe81Cm5hR+uKQCKcXoSrn1UO7OqF0YGytKQFSc3cOYwlee2 vNSUXFq2QbBDoYqs34KisUVJ7M2V/V72xkTUDUQKilF6isZOg472e95RvmE5UkeqMkQa KI//HCdSt852EFeOgNl1SIdrxwVuQ18KmItwEv9Pf0ab9gw5QQaLm/j3V5UlMh2o3Dz1 1xlA== X-Forwarded-Encrypted: i=1; AJvYcCVfze6YFJpDOW0xQfJOdTqV9lSMwxgKQizJdid2OJP/FQND3EEFq+eMFocofmKK5OQcO4oS8jYB+0X0tSs=@vger.kernel.org X-Gm-Message-State: AOJu0YxAyvT2v3y+kodhm8aD+nDI7yhn0k3Uxg6VROAONppZR+/ynrMo DkMWqIpLATSLscE+ZhOMDPir5pGi9cBTeUCcPGMX8DkIk8KAcRNy7dvjG6kkzv11D0lzIMssJ4K 7PhtIsU7DbA== X-Google-Smtp-Source: AGHT+IEW+7hRNxjkqclt5Tj3p4dCnP73X9oW+uaLcqCDwWscrSXsDam/MnTmwIG6QENAU+LIBRRdzWl34CiC X-Received: from plnu14.prod.google.com ([2002:a17:903:1a4e:b0:240:7c7d:49db]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:902:ecc3:b0:244:214f:13a0 with SMTP id d9443c01a7336-2447909b787mr147077165ad.52.1755543971301; Mon, 18 Aug 2025 12:06:11 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:11 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-16-irogers@google.com> Subject: [PATCH v2 15/20] perf vendor events: Update rocketlake metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/rocketlake/rkl-metrics.json | 97 +++++++++++-------- 1 file changed, 56 insertions(+), 41 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/rocketlake/rkl-metrics.json b/t= ools/perf/pmu-events/arch/x86/rocketlake/rkl-metrics.json index 71737a1a1997..79dc34157481 100644 --- a/tools/perf/pmu-events/arch/x86/rocketlake/rkl-metrics.json +++ b/tools/perf/pmu-events/arch/x86/rocketlake/rkl-metrics.json @@ -1,63 +1,63 @@ [ { "BriefDescription": "C10 residency percent per package", - "MetricExpr": "cstate_pkg@c10\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c10\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C10_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C8 residency percent per package", - "MetricExpr": "cstate_pkg@c8\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c8\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C8_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C9 residency percent per package", - "MetricExpr": "cstate_pkg@c9\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c9\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C9_Pkg_Residency", "ScaleUnit": "100%" @@ -85,7 +85,6 @@ }, { "BriefDescription": "This metric estimates how often memory load a= ccesses were aliased by preceding stores (in program order) with a 4K addre= ss offset", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "LD_BLOCKS_PARTIAL.ADDRESS_ALIAS / tma_info_thread_c= lks", "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", "MetricName": "tma_4k_aliasing", @@ -134,6 +133,7 @@ }, { "BriefDescription": "Total pipeline cost of instruction fetch rela= ted bottlenecks by large code footprint programs (i-side cache; TLB and BTB= misses)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * tma_fetch_latency * (tma_itlb_misses + tma_ic= ache_misses + tma_unknown_branches) / (tma_branch_resteers + tma_dsb_switch= es + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches)", "MetricGroup": "BigFootprint;BvBC;Fed;Frontend;IcMiss;MemoryTLB", "MetricName": "tma_bottleneck_big_code", @@ -147,40 +147,45 @@ "MetricThreshold": "tma_bottleneck_branching_overhead > 5", "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_4k_aliasing + tma_dtlb_load + tma_= fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + = tma_store_fwd_blk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l= 1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_b= lk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + = tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_= 4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma= _lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * = (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_boun= d + tma_store_bound)) * (tma_split_loads / (tma_4k_aliasing + tma_dtlb_load= + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_l= oads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_dra= m_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * = (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency + tma_streaming_stores)) + tma_memory_bound * (tma_stor= e_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tm= a_store_bound)) * (tma_store_latency / (tma_dtlb_store + tma_false_sharing = + tma_split_stores + tma_store_latency + tma_streaming_stores)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + 10 * tma_microcode_seq= uencer * tma_other_mispredicts / tma_branch_mispredicts * tma_mispredicts_r= esteers) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_br= anches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tm= a_itlb_misses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * tma_ms /= (tma_dsb + tma_lsd + tma_mite + tma_ms))) - tma_bottleneck_big_code", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + 10 * tma_microcode_seq= uencer * tma_other_mispredicts / tma_branch_mispredicts * tma_mispredicts_r= esteers) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_br= anches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tm= a_itlb_misses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottleneck_big_= code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20" }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cle= ars_resteers + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_b= ranch_mispredicts * tma_mispredicts_resteers) / (tma_clears_resteers + tma_= mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_= dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switc= hes) + tma_fetch_bandwidth * tma_ms / (tma_dsb + tma_lsd + tma_mite + tma_m= s)) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_mis= predicts * tma_branch_mispredicts + tma_machine_clears * tma_other_nukes / = tma_other_nukes + tma_core_bound * (tma_serializing_operation + tma_core_bo= und * RS_EVENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_utilized_0)= / (tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_= microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer)= * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cle= ars_resteers + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_b= ranch_mispredicts * tma_mispredicts_resteers) / (tma_clears_resteers + tma_= mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_= dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switc= hes) + tma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma= _branch_mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_oth= er_nukes / tma_other_nukes + tma_core_bound * (tma_serializing_operation + = tma_core_bound * RS_EVENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_= utilized_0) / (tma_divider + tma_ports_utilization + tma_serializing_operat= ion) + tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode= _sequencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operation= s)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -188,6 +193,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + = tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, tma_4k_aliasing + tm= a_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + = tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound= / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store= _bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_spli= t_stores + tma_store_latency + tma_streaming_stores)))", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", @@ -196,6 +202,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Synchronization= related bottlenecks (data transfers and coherency updates across processor= s)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_l3_bound / (tma_dram= _bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * (t= ma_contested_accesses + tma_data_sharing) / (tma_contested_accesses + tma_d= ata_sharing + tma_l3_hit_latency + tma_sq_full) + tma_store_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * = tma_false_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency + tma_streaming_stores - tma_store_latency)) + tma_mach= ine_clears * (1 - tma_other_nukes / tma_other_nukes))", "MetricGroup": "BvMS;LockCont;Mem;Offcore;tma_issueSyncxn", "MetricName": "tma_bottleneck_memory_synchronization", @@ -204,6 +211,7 @@ }, { "BriefDescription": "Total pipeline cost of Branch Misprediction r= elated bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (1 - 10 * tma_microcode_sequencer * tma_other= _mispredicts / tma_branch_mispredicts) * (tma_branch_mispredicts + tma_fetc= h_latency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switc= hes + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches))", "MetricGroup": "Bad;BadSpec;BrMispredicts;BvMP;tma_issueBM", "MetricName": "tma_bottleneck_mispredictions", @@ -212,7 +220,8 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -220,6 +229,7 @@ }, { "BriefDescription": "Total pipeline cost of \"useful operations\" = - the portion of Retiring category not covered by Branching_Overhead nor Ir= regular_Overhead.", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_retiring - (BR_INST_RETIRED.ALL_BRANCHES= + 2 * BR_INST_RETIRED.NEAR_CALL + INST_RETIRED.NOP) / tma_info_thread_slot= s - tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_se= quencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "BvUW;Ret", "MetricName": "tma_bottleneck_useful_work", @@ -427,7 +437,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%" }, { @@ -619,6 +629,7 @@ }, { "BriefDescription": "Total pipeline cost of DSB (uop cache) hits -= subset of the Instruction_Fetch_BW Bottleneck", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_frontend_bound * (tma_fetch_bandwidth / = (tma_fetch_bandwidth + tma_fetch_latency)) * (tma_dsb / (tma_dsb + tma_lsd = + tma_mite + tma_ms)))", "MetricGroup": "DSB;Fed;FetchBW;tma_issueFB", "MetricName": "tma_info_botlnk_l2_dsb_bandwidth", @@ -1074,7 +1085,7 @@ "MetricName": "tma_info_memory_tlb_store_stlb_mpki" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (UOPS_EXECUTED.CORE_CYCLES_G= E_1 / 2 if #SMT_on else cpu@UOPS_EXECUTED.THREAD\\,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -1097,6 +1108,12 @@ "MetricGroup": "Fed;FetchBW", "MetricName": "tma_info_pipeline_fetch_mite" }, + { + "BriefDescription": "Average number of uops fetched from MS per cy= cle", + "MetricExpr": "IDQ.MS_UOPS / cpu@IDQ.MS_UOPS\\,cmask\\=3D1@", + "MetricGroup": "Fed;FetchLat;MicroSeq", + "MetricName": "tma_info_pipeline_fetch_ms" + }, { "BriefDescription": "Instructions per a microcode Assist invocatio= n", "MetricExpr": "INST_RETIRED.ANY / ASSISTS.ANY", @@ -1113,7 +1130,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -1125,7 +1142,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -1134,7 +1151,7 @@ "MetricExpr": "64 * (UNC_ARB_TRK_REQUESTS.ALL + UNC_ARB_COH_TRK_RE= QUESTS.ALL) / 1e6 / tma_info_system_time / 1e3", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Giga Floating Point Operations Per Second", @@ -1165,6 +1182,7 @@ }, { "BriefDescription": "Average number of parallel data read requests= to external memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "UNC_ARB_DAT_OCCUPANCY.RD / UNC_ARB_DAT_OCCUPANCY.RD= @cmask\\=3D1@", "MetricGroup": "Mem;MemoryBW;SoC", "MetricName": "tma_info_system_mem_parallel_reads", @@ -1316,12 +1334,12 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", "MetricExpr": "min(2 * (MEM_INST_RETIRED.ALL_LOADS - MEM_LOAD_RETI= RED.FB_HIT - MEM_LOAD_RETIRED.L1_MISS) * 20 / 100, max(CYCLE_ACTIVITY.CYCLE= S_MEM_ANY - CYCLE_ACTIVITY.CYCLES_L1D_MISS, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%" }, { @@ -1345,7 +1363,6 @@ }, { "BriefDescription": "This metric estimates how often the CPU was s= talled due to loads accesses to L3 cache or contended with a sibling Core", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(CYCLE_ACTIVITY.STALLS_L2_MISS - CYCLE_ACTIVITY.STA= LLS_L3_MISS) / tma_info_thread_clks", "MetricGroup": "CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_gr= oup;tma_memory_bound_group", "MetricName": "tma_l3_bound", @@ -1359,7 +1376,7 @@ "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%" }, { @@ -1465,7 +1482,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%" }, { @@ -1474,7 +1491,7 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%" }, { @@ -1542,7 +1559,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "cpu@IDQ.MS_UOPS\\,cmask\\=3D1@ / tma_info_core_core= _clks / 2", + "MetricExpr": "cpu@IDQ.MS_UOPS\\,cmask\\=3D1@ / tma_info_core_core= _clks / 3.3", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -1676,7 +1693,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_info_= thread_slots", + "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -1713,7 +1730,6 @@ }, { "BriefDescription": "This metric represents rate of split store ac= cesses", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "MEM_INST_RETIRED.SPLIT_STORES / tma_info_core_core_= clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_issueSpSt;tma_store_bou= nd_group", "MetricName": "tma_split_stores", @@ -1727,7 +1743,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%" }, { @@ -1741,7 +1757,6 @@ }, { "BriefDescription": "This metric roughly estimates fraction of cyc= les when the memory subsystem had loads blocked since they could not forwar= d data from earlier (in program order) overlapping stores", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * LD_BLOCKS.STORE_FORWARD / tma_info_thread_clks= ", "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", "MetricName": "tma_store_fwd_blk", --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pl1-f202.google.com (mail-pl1-f202.google.com [209.85.214.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 05962273D8F for ; Mon, 18 Aug 2025 19:06:13 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.202 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543975; cv=none; b=rh3NfloEeY9bRtbFw/iV5veyM90ui6lqFzO5qFYluoOBw4lj69pHRsZ7mjNl/oZ3NR5Qv7L0E9s2QhVAD631n+ezjqvH48mPUZXk++gn2ESMox1X589dM7OsTmbXAQleeGZWK07WSe7XPWtHaxkPMwBuwNlxyRzbLVYasbcI6Ik= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543975; c=relaxed/simple; bh=LcmzKh8MxKe7S2JdMF2i56oqJa5WycnupoSJN/TemBI=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=jEYWQcos+c4r9k3X8RACoF++VifDLbQBNDPzKCH/tEygkcWeqTZyACgm++GwLSnz+8BQXCuW2M1Y8vlvhuK5TMiTb3iVjYE8QjXSfwYDo9uOHHlaWIG3YGpgRAu8XAOiAyvEqL6nPvIn5S5ogbZ6tBvyEchEmFcsUYx5BgMzaY8= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=zHjzcCKT; arc=none smtp.client-ip=209.85.214.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="zHjzcCKT" Received: by mail-pl1-f202.google.com with SMTP id d9443c01a7336-24457fe8844so51612015ad.0 for ; Mon, 18 Aug 2025 12:06:13 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543973; x=1756148773; darn=vger.kernel.org; h=to:from:subject:message-id:references:mime-version:in-reply-to:date :from:to:cc:subject:date:message-id:reply-to; bh=1Oh45/+Kol4/Ff6MHQPrzumdu5r/WMLLUjenm77C2bc=; b=zHjzcCKTge2ZI8k73TEuNUVWI7U+zOzK2CkEfJ4GgZeayd4UVmQjhdChD12Lko8aHV e4d61tPdf+bG0S1Zx5+UDj024+OElfV8LgQ6QFNVQtOIQGjPz54+VVz+hT2cwmsyDfYR L+ZPtR+h3/ro8aLcu2dsxnz1VyqYSX0AaiOBS3IDqwXP3sjrgSTAmJIgAxEC2o44Usd9 iwhELy/SefCqBENbI7fII/1KACNfVhG1AUbT4Ujs7D9zEsyVP4dU283UcJbrn5KcFE+g h0RHepgzmaRs2+diy2YCM62LCct9LalVYjWf1k+FflBsuoORj2ggiUG2j11lnwowyJik wtaQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543973; x=1756148773; h=to:from:subject:message-id:references:mime-version:in-reply-to:date :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=1Oh45/+Kol4/Ff6MHQPrzumdu5r/WMLLUjenm77C2bc=; b=HfJ5FG/zbrnjnAuopmZ5h3uRT2182wje2t0b+jjZBPReFhk5EirGFSivVBeagHv7BX pDwNrHsZITTU4rzrHylQ7dFCLFTGTCsoWteiWQWNA/y9XGzGtJwwRXrE2GGJNbTCnB8y FdBKl57BhvvzxgseLHfceR8S+M4T/SudrLUcDlQAH4ogV0NUSO/NEXOAcnGuiY9L3+7b cch1EDjAwSQzmFBD7jLyKwdv4Ig45jPXafQ/nASuVCwOD4NhU/GrxvDrSvxsinNVuCG5 fsdECWRfOIai3R3WOFpu97a8qglzKZ7URA9/T3DVYA01lFuc7x6rsYh6jTz5ZeUVnXWP HOUg== X-Forwarded-Encrypted: i=1; AJvYcCXGWh426b38Gp60papBI9nc0m/MiqWY9PIdg4dAkrjc0AVyOZOFqEueUt81qMR1wI/pffCDozZnwmnhiiY=@vger.kernel.org X-Gm-Message-State: AOJu0YxBzOfEEdka3tiJZLc9OHTJuEo1qO2lujI0OKTuy0BBPHi7q5GP lzPPGJvIBDSs4XnwRuRZzj8rkzGG6i/k7tw1oDeZ5gW+LWGMIOvi5ENgBsx3nkMNkRU1tk8kZCU Uvrse6oKXGg== X-Google-Smtp-Source: AGHT+IHTGaTTiW/uZLksoxGLFmmaMCpk22DofrBNoAabsB3kcFxlVjkcIJyuzq08E+2twTPTPd3iYeDDlNTi X-Received: from pjbqx14.prod.google.com ([2002:a17:90b:3e4e:b0:321:9285:be8]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:90b:1f83:b0:321:87fa:e1f6 with SMTP id 98e67ed59e1d1-3245e562016mr15244a91.1.1755543973327; Mon, 18 Aug 2025 12:06:13 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:12 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-17-irogers@google.com> Subject: [PATCH v2 16/20] perf vendor events: Update sandybridge metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/sandybridge/snb-metrics.json | 19 +++++++++---------- 1 file changed, 9 insertions(+), 10 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/sandybridge/snb-metrics.json b/= tools/perf/pmu-events/arch/x86/sandybridge/snb-metrics.json index 823d8b7c4224..d40761903429 100644 --- a/tools/perf/pmu-events/arch/x86/sandybridge/snb-metrics.json +++ b/tools/perf/pmu-events/arch/x86/sandybridge/snb-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -71,7 +71,6 @@ }, { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "1 - (tma_frontend_bound + tma_bad_speculation + tma= _retiring)", "MetricGroup": "BvOB;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", @@ -296,7 +295,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -308,7 +307,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-qk1-f202.google.com (mail-qk1-f202.google.com [209.85.222.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 3456132C33A for ; Mon, 18 Aug 2025 19:06:16 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.222.202 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543991; cv=none; b=FyHJb/qb8I9sDxVEqi/1CsClGlZPxu9HYDkqc6ZSB082uw2ry2I8jjDYSD8DRvyPVSexGBVkmsBgA+Vn+pD853TEeaXCMUPygL6KEG/W3maJVk0Eu4dpSyfgtIKXuBJeY0GC+Xa+0nR4qu8EujG2zzv/k6Xu5N1fhiL+AVrLnMQ= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543991; c=relaxed/simple; bh=pDxmFR4AM0/Vu4fAtJ2/cf5EWtDlozhxI61S+Ah/wa0=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=GarSeLKSw0OHxER5ZANsFx/H4nvfDmYOj2qUeMl9TJwZ5MD0vADW1/k/SFyOgo8gOMXHrkhvyyIpw/TdmNyeihh2p6arbJ+inOXVOd04ZguiEkqh12T77svNADGSpsNdNh5J2CiXMN50w3QRDHSs0jZ/TxspMqpAYgwcgKkOZWQ= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=mpgHJ04v; arc=none smtp.client-ip=209.85.222.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="mpgHJ04v" Received: by mail-qk1-f202.google.com with SMTP id af79cd13be357-7e87062832bso1353344685a.2 for ; Mon, 18 Aug 2025 12:06:16 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543976; x=1756148776; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=DoOMUNx7yVNc4GFT6CFi+NjihGSZ7ZZhLQ1fsm/fhLI=; b=mpgHJ04vhc2tRFKMKHAAIQ6qLfFG7lxYWo2a6PlHZLZFARsSE9oj6asaV5EVkyRrrL F18PI2MnS6rH7g25/LwUgv5EyX3ShzRSW4B+uFFsO6owk4ykFW6CxIJF+Pc14VUsXvye poRyZ0RZXOJVZ0q4PLy2t5ijAk6M4NmX27ZDbVBjtQy1Pt3t+4SMH7b46WgVHnCCn1Sk qFUuXiu1bp+4QcRGtfTZcu/cXO8MKKBtv9NXpFwm6EoYXrJSnD9ogvrT25mH/6Trk1Ql MtNstb8xS5JS+YPEvBmuEcncty5mhs6I20I1Y1Ckvb2o9QXNPY5mn6pB6o12sl3HbVB8 Wchg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543976; x=1756148776; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=DoOMUNx7yVNc4GFT6CFi+NjihGSZ7ZZhLQ1fsm/fhLI=; b=UQK+TZDVpQ5MhABy1DlTJ29QfZq1cTn391uQStMZ+ejPfTS+BxX2Z24mc9kfj9nWRi aan5L8by0w2Bzn78prD/9RMG0yai8z8fJGLUn5z3/Ci+I3UrzRFFoU/+L3CuJ+ldirt/ 1UIPsofwaO7d/G2ItVpKyAkdUojmQt35S1a1j91dz5loTAItVkq4zw78vWTEvpIO5dwi esdzLcBjJO7FqMtRsalwl4197REtAYQY79U0v53Uue/5+liKW8t0rjnLerg1K/zSlLGp OJGyijYUG1GpDa2oeGRLFN3LiGS9U1xSyFmRByPSzoFMHiBSKaeWbYw9ONxxtCrX4ZiM DNRA== X-Forwarded-Encrypted: i=1; AJvYcCVE011xhcajfuLVJ63CYR2z4IP/FB/+sQG8nao3tHXg/QZE0Og5QwugDoeadHefjF700bkGzeyX1ap8F24=@vger.kernel.org X-Gm-Message-State: AOJu0YxMo7aaw2gRr8UFCVEg/Mr9m1l1jLCOak8xrLJn9tio3tLiWBhG K0NObVmG7X3tacSeDJzPZ/Hr0GrL6ZSOYqgmexRxPsuwAKHZp/tck3OzJCQ1yrI9eGz1w75YmwE SFCLkD7KB5w== X-Google-Smtp-Source: AGHT+IHayDp9rqv3JQqGGkjdoJKyKdyLCrPNwKmVR03JbQM6+0wY1YgTzZJIBB2AVeh0TJVGqQ39haTqjp0d X-Received: from qkpi22.prod.google.com ([2002:a05:620a:27d6:b0:7e8:336:737b]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a05:620a:28d4:b0:7e6:6ea5:aae2 with SMTP id af79cd13be357-7e9f2bcc2edmr43402685a.11.1755543975452; Mon, 18 Aug 2025 12:06:15 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:13 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-18-irogers@google.com> Subject: [PATCH v2 17/20] perf vendor events: Update sapphirerapids events/metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update events from v1.28 to v1.30. Update metrics from TMA 5.0 to 5.1. The event updates come from: https://github.com/intel/perfmon/commit/c6a01e651c7be0dbc7a0e92ea915bb3c7e5= 970da https://github.com/intel/perfmon/commit/8b3a5b3f8ebf3cc48e29e3b65ecccb37f6f= c3e81 Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- tools/perf/pmu-events/arch/x86/mapfile.csv | 2 +- .../arch/x86/sapphirerapids/cache.json | 100 +++++------ .../x86/sapphirerapids/floating-point.json | 43 ++--- .../arch/x86/sapphirerapids/frontend.json | 42 +++-- .../arch/x86/sapphirerapids/memory.json | 30 ++-- .../arch/x86/sapphirerapids/other.json | 28 ++- .../arch/x86/sapphirerapids/pipeline.json | 167 +++++++----------- .../arch/x86/sapphirerapids/spr-metrics.json | 153 ++++++++++------ .../x86/sapphirerapids/uncore-memory.json | 82 +++++++++ .../x86/sapphirerapids/virtual-memory.json | 40 ++--- 10 files changed, 382 insertions(+), 305 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/mapfile.csv b/tools/perf/pmu-ev= ents/arch/x86/mapfile.csv index f8bf16d30602..d9daab4d8461 100644 --- a/tools/perf/pmu-events/arch/x86/mapfile.csv +++ b/tools/perf/pmu-events/arch/x86/mapfile.csv @@ -29,7 +29,7 @@ GenuineIntel-6-2E,v4,nehalemex,core GenuineIntel-6-CC,v1.00,pantherlake,core GenuineIntel-6-A7,v1.04,rocketlake,core GenuineIntel-6-2A,v19,sandybridge,core -GenuineIntel-6-8F,v1.28,sapphirerapids,core +GenuineIntel-6-8F,v1.30,sapphirerapids,core GenuineIntel-6-AF,v1.11,sierraforest,core GenuineIntel-6-(37|4A|4C|4D|5A),v15,silvermont,core GenuineIntel-6-(4E|5E|8E|9E|A5|A6),v59,skylake,core diff --git a/tools/perf/pmu-events/arch/x86/sapphirerapids/cache.json b/too= ls/perf/pmu-events/arch/x86/sapphirerapids/cache.json index 21db53f9e9d6..a30af5140e6c 100644 --- a/tools/perf/pmu-events/arch/x86/sapphirerapids/cache.json +++ b/tools/perf/pmu-events/arch/x86/sapphirerapids/cache.json @@ -4,7 +4,6 @@ "Counter": "0,1,2,3", "EventCode": "0x51", "EventName": "L1D.HWPF_MISS", - "PublicDescription": "L1D.HWPF_MISS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -13,7 +12,7 @@ "Counter": "0,1,2,3", "EventCode": "0x51", "EventName": "L1D.REPLACEMENT", - "PublicDescription": "Counts L1D data line replacements including = opportunistic replacements, and replacements that require stall-for-replace= or block-for-replace. Available PDIST counters: 0", + "PublicDescription": "Counts L1D data line replacements including = opportunistic replacements, and replacements that require stall-for-replace= or block-for-replace.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -22,7 +21,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.FB_FULL", - "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -33,7 +32,7 @@ "EdgeDetect": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.FB_FULL_PERIODS", - "PublicDescription": "Counts number of phases a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses. Av= ailable PDIST counters: 0", + "PublicDescription": "Counts number of phases a demand request has= waited due to L1D Fill Buffer (FB) unavailability. Demand requests include= cacheable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -43,7 +42,6 @@ "Deprecated": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.L2_STALL", - "PublicDescription": "This event is deprecated. Refer to new event= L1D_PEND_MISS.L2_STALLS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -52,7 +50,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.L2_STALLS", - "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D due to lack of L2 resources. Demand requests include cac= heable/uncacheable demand load, store, lock or SW prefetch accesses. Availa= ble PDIST counters: 0", + "PublicDescription": "Counts number of cycles a demand request has= waited due to L1D due to lack of L2 resources. Demand requests include cac= heable/uncacheable demand load, store, lock or SW prefetch accesses.", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -61,7 +59,7 @@ "Counter": "0,1,2,3", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.PENDING", - "PublicDescription": "Counts number of L1D misses that are outstan= ding in each cycle, that is each cycle the number of Fill Buffers (FB) outs= tanding required by Demand Reads. FB either is held by demand loads, or it = is held by non-demand loads and gets hit at least once by demand. The valid= outstanding interval is defined until the FB deallocation by one of the fo= llowing ways: from FB allocation, if FB is allocated by demand from the dem= and Hit FB, if it is allocated by hardware or software prefetch. Note: In t= he L1D, a Demand Read contains cacheable or noncacheable demand loads, incl= uding ones causing cache-line splits and reads due to page walks resulted f= rom any request type. Available PDIST counters: 0", + "PublicDescription": "Counts number of L1D misses that are outstan= ding in each cycle, that is each cycle the number of Fill Buffers (FB) outs= tanding required by Demand Reads. FB either is held by demand loads, or it = is held by non-demand loads and gets hit at least once by demand. The valid= outstanding interval is defined until the FB deallocation by one of the fo= llowing ways: from FB allocation, if FB is allocated by demand from the dem= and Hit FB, if it is allocated by hardware or software prefetch. Note: In t= he L1D, a Demand Read contains cacheable or noncacheable demand loads, incl= uding ones causing cache-line splits and reads due to page walks resulted f= rom any request type.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -71,7 +69,7 @@ "CounterMask": "1", "EventCode": "0x48", "EventName": "L1D_PEND_MISS.PENDING_CYCLES", - "PublicDescription": "Counts duration of L1D miss outstanding in c= ycles. Available PDIST counters: 0", + "PublicDescription": "Counts duration of L1D miss outstanding in c= ycles.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -80,7 +78,7 @@ "Counter": "0,1,2,3", "EventCode": "0x25", "EventName": "L2_LINES_IN.ALL", - "PublicDescription": "Counts the number of L2 cache lines filling = the L2. Counting does not cover rejects. Available PDIST counters: 0", + "PublicDescription": "Counts the number of L2 cache lines filling = the L2. Counting does not cover rejects.", "SampleAfterValue": "100003", "UMask": "0x1f" }, @@ -89,7 +87,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.NON_SILENT", - "PublicDescription": "Counts the number of lines that are evicted = by L2 cache when triggered by an L2 cache fill. Those lines are in Modified= state. Modified lines are written back to L3 Available PDIST counters: 0", + "PublicDescription": "Counts the number of lines that are evicted = by L2 cache when triggered by an L2 cache fill. Those lines are in Modified= state. Modified lines are written back to L3", "SampleAfterValue": "200003", "UMask": "0x2" }, @@ -98,7 +96,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.SILENT", - "PublicDescription": "Counts the number of lines that are silently= dropped by L2 cache. These lines are typically in Shared or Exclusive stat= e. A non-threaded event. Available PDIST counters: 0", + "PublicDescription": "Counts the number of lines that are silently= dropped by L2 cache. These lines are typically in Shared or Exclusive stat= e. A non-threaded event.", "SampleAfterValue": "200003", "UMask": "0x1" }, @@ -107,7 +105,7 @@ "Counter": "0,1,2,3", "EventCode": "0x26", "EventName": "L2_LINES_OUT.USELESS_HWPF", - "PublicDescription": "Counts the number of cache lines that have b= een prefetched by the L2 hardware prefetcher but not used by demand access = when evicted from the L2 cache Available PDIST counters: 0", + "PublicDescription": "Counts the number of cache lines that have b= een prefetched by the L2 hardware prefetcher but not used by demand access = when evicted from the L2 cache", "SampleAfterValue": "200003", "UMask": "0x4" }, @@ -116,7 +114,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.ALL", - "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_RQSTS.REFERENCES] Available PDIST coun= ters: 0", + "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_RQSTS.REFERENCES]", "SampleAfterValue": "200003", "UMask": "0xff" }, @@ -125,7 +123,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_REQUEST.MISS", - "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_RQSTS.MISS] Available PDIST coun= ters: 0", + "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_RQSTS.MISS]", "SampleAfterValue": "200003", "UMask": "0x3f" }, @@ -134,7 +132,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_CODE_RD", - "PublicDescription": "Counts the total number of L2 code requests.= Available PDIST counters: 0", + "PublicDescription": "Counts the total number of L2 code requests.= ", "SampleAfterValue": "200003", "UMask": "0xe4" }, @@ -143,7 +141,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_DATA_RD", - "PublicDescription": "Counts Demand Data Read requests accessing t= he L2 cache. These requests may hit or miss L2 cache. True-miss exclude mis= ses that were merged with ongoing L2 misses. An access is counted once. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts Demand Data Read requests accessing t= he L2 cache. These requests may hit or miss L2 cache. True-miss exclude mis= ses that were merged with ongoing L2 misses. An access is counted once.", "SampleAfterValue": "200003", "UMask": "0xe1" }, @@ -152,7 +150,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_MISS", - "PublicDescription": "Counts demand requests that miss L2 cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts demand requests that miss L2 cache.", "SampleAfterValue": "200003", "UMask": "0x27" }, @@ -161,7 +159,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_DEMAND_REFERENCES", - "PublicDescription": "Counts demand requests to L2 cache. Availabl= e PDIST counters: 0", + "PublicDescription": "Counts demand requests to L2 cache.", "SampleAfterValue": "200003", "UMask": "0xe7" }, @@ -170,7 +168,6 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_HWPF", - "PublicDescription": "L2_RQSTS.ALL_HWPF Available PDIST counters: = 0", "SampleAfterValue": "200003", "UMask": "0xf0" }, @@ -179,7 +176,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.ALL_RFO", - "PublicDescription": "Counts the total number of RFO (read for own= ership) requests to L2 cache. L2 RFO requests include both L1D demand RFO m= isses as well as L1D RFO prefetches. Available PDIST counters: 0", + "PublicDescription": "Counts the total number of RFO (read for own= ership) requests to L2 cache. L2 RFO requests include both L1D demand RFO m= isses as well as L1D RFO prefetches.", "SampleAfterValue": "200003", "UMask": "0xe2" }, @@ -188,7 +185,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.CODE_RD_HIT", - "PublicDescription": "Counts L2 cache hits when fetching instructi= ons, code reads. Available PDIST counters: 0", + "PublicDescription": "Counts L2 cache hits when fetching instructi= ons, code reads.", "SampleAfterValue": "200003", "UMask": "0xc4" }, @@ -197,7 +194,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.CODE_RD_MISS", - "PublicDescription": "Counts L2 cache misses when fetching instruc= tions. Available PDIST counters: 0", + "PublicDescription": "Counts L2 cache misses when fetching instruc= tions.", "SampleAfterValue": "200003", "UMask": "0x24" }, @@ -206,7 +203,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT", - "PublicDescription": "Counts the number of demand Data Read reques= ts initiated by load instructions that hit L2 cache. Available PDIST counte= rs: 0", + "PublicDescription": "Counts the number of demand Data Read reques= ts initiated by load instructions that hit L2 cache.", "SampleAfterValue": "200003", "UMask": "0xc1" }, @@ -215,7 +212,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.DEMAND_DATA_RD_MISS", - "PublicDescription": "Counts demand Data Read requests with true-m= iss in the L2 cache. True-miss excludes misses that were merged with ongoin= g L2 misses. An access is counted once. Available PDIST counters: 0", + "PublicDescription": "Counts demand Data Read requests with true-m= iss in the L2 cache. True-miss excludes misses that were merged with ongoin= g L2 misses. An access is counted once.", "SampleAfterValue": "200003", "UMask": "0x21" }, @@ -224,7 +221,6 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.HWPF_MISS", - "PublicDescription": "L2_RQSTS.HWPF_MISS Available PDIST counters:= 0", "SampleAfterValue": "200003", "UMask": "0x30" }, @@ -233,7 +229,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.MISS", - "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_REQUEST.MISS] Available PDIST co= unters: 0", + "PublicDescription": "Counts read requests of any type with true-m= iss in the L2 cache. True-miss excludes L2 misses that were merged with ong= oing L2 misses. [This event is alias to L2_REQUEST.MISS]", "SampleAfterValue": "200003", "UMask": "0x3f" }, @@ -242,7 +238,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.REFERENCES", - "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_REQUEST.ALL] Available PDIST counters:= 0", + "PublicDescription": "Counts all requests that were hit or true mi= sses in L2 cache. True-miss excludes misses that were merged with ongoing L= 2 misses. [This event is alias to L2_REQUEST.ALL]", "SampleAfterValue": "200003", "UMask": "0xff" }, @@ -251,7 +247,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.RFO_HIT", - "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that hit L2 cache. Available PDIST counters: 0", + "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that hit L2 cache.", "SampleAfterValue": "200003", "UMask": "0xc2" }, @@ -260,7 +256,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.RFO_MISS", - "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that miss L2 cache. Available PDIST counters: 0", + "PublicDescription": "Counts the RFO (Read-for-Ownership) requests= that miss L2 cache.", "SampleAfterValue": "200003", "UMask": "0x22" }, @@ -269,7 +265,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.SWPF_HIT", - "PublicDescription": "Counts Software prefetch requests that hit t= he L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when = FB is not full. Available PDIST counters: 0", + "PublicDescription": "Counts Software prefetch requests that hit t= he L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when = FB is not full.", "SampleAfterValue": "200003", "UMask": "0xc8" }, @@ -278,7 +274,7 @@ "Counter": "0,1,2,3", "EventCode": "0x24", "EventName": "L2_RQSTS.SWPF_MISS", - "PublicDescription": "Counts Software prefetch requests that miss = the L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when= FB is not full. Available PDIST counters: 0", + "PublicDescription": "Counts Software prefetch requests that miss = the L2 cache. Accounts for PREFETCHNTA and PREFETCHT0/1/2 instructions when= FB is not full.", "SampleAfterValue": "200003", "UMask": "0x28" }, @@ -287,7 +283,7 @@ "Counter": "0,1,2,3", "EventCode": "0x23", "EventName": "L2_TRANS.L2_WB", - "PublicDescription": "Counts L2 writebacks that access L2 cache. A= vailable PDIST counters: 0", + "PublicDescription": "Counts L2 writebacks that access L2 cache.", "SampleAfterValue": "200003", "UMask": "0x40" }, @@ -296,7 +292,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x2e", "EventName": "LONGEST_LAT_CACHE.MISS", - "PublicDescription": "Counts core-originated cacheable requests th= at miss the L3 cache (Longest Latency cache). Requests include data and cod= e reads, Reads-for-Ownership (RFOs), speculative accesses and hardware pref= etches to the L1 and L2. It does not include hardware prefetches to the L3= , and may not count other types of requests to the L3. Available PDIST coun= ters: 0", + "PublicDescription": "Counts core-originated cacheable requests th= at miss the L3 cache (Longest Latency cache). Requests include data and cod= e reads, Reads-for-Ownership (RFOs), speculative accesses and hardware pref= etches to the L1 and L2. It does not include hardware prefetches to the L3= , and may not count other types of requests to the L3.", "SampleAfterValue": "100003", "UMask": "0x41" }, @@ -305,7 +301,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x2e", "EventName": "LONGEST_LAT_CACHE.REFERENCE", - "PublicDescription": "Counts core-originated cacheable requests to= the L3 cache (Longest Latency cache). Requests include data and code reads= , Reads-for-Ownership (RFOs), speculative accesses and hardware prefetches = to the L1 and L2. It does not include hardware prefetches to the L3, and m= ay not count other types of requests to the L3. Available PDIST counters: 0= ", + "PublicDescription": "Counts core-originated cacheable requests to= the L3 cache (Longest Latency cache). Requests include data and code reads= , Reads-for-Ownership (RFOs), speculative accesses and hardware prefetches = to the L1 and L2. It does not include hardware prefetches to the L3, and m= ay not count other types of requests to the L3.", "SampleAfterValue": "100003", "UMask": "0x4f" }, @@ -394,7 +390,7 @@ "Counter": "0,1,2,3", "EventCode": "0x43", "EventName": "MEM_LOAD_COMPLETED.L1_MISS_ANY", - "PublicDescription": "Number of completed demand load requests tha= t missed the L1 data cache including shadow misses (FB hits, merge to an on= going L1D miss) Available PDIST counters: 0", + "PublicDescription": "Number of completed demand load requests tha= t missed the L1 data cache including shadow misses (FB hits, merge to an on= going L1D miss)", "SampleAfterValue": "1000003", "UMask": "0xfd" }, @@ -582,7 +578,6 @@ "Counter": "0,1,2,3", "EventCode": "0x44", "EventName": "MEM_STORE_RETIRED.L2_HIT", - "PublicDescription": "MEM_STORE_RETIRED.L2_HIT Available PDIST cou= nters: 0", "SampleAfterValue": "200003", "UMask": "0x1" }, @@ -591,7 +586,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe5", "EventName": "MEM_UOP_RETIRED.ANY", - "PublicDescription": "Number of retired micro-operations (uops) fo= r load or store memory accesses Available PDIST counters: 0", + "PublicDescription": "Number of retired micro-operations (uops) fo= r load or store memory accesses", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -1073,7 +1068,6 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.ALL_REQUESTS", - "PublicDescription": "OFFCORE_REQUESTS.ALL_REQUESTS Available PDIS= T counters: 0", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -1082,7 +1076,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DATA_RD", - "PublicDescription": "Counts the demand and prefetch data reads. A= ll Core Data Reads include cacheable 'Demands' and L2 prefetchers (not L3 p= refetchers). Counting also covers reads due to page walks resulted from any= request type. Available PDIST counters: 0", + "PublicDescription": "Counts the demand and prefetch data reads. A= ll Core Data Reads include cacheable 'Demands' and L2 prefetchers (not L3 p= refetchers). Counting also covers reads due to page walks resulted from any= request type.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -1091,7 +1085,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_CODE_RD", - "PublicDescription": "Counts both cacheable and non-cacheable code= read requests. Available PDIST counters: 0", + "PublicDescription": "Counts both cacheable and non-cacheable code= read requests.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -1100,7 +1094,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_DATA_RD", - "PublicDescription": "Counts the Demand Data Read requests sent to= uncore. Use it in conjunction with OFFCORE_REQUESTS_OUTSTANDING to determi= ne average latency in the uncore. Available PDIST counters: 0", + "PublicDescription": "Counts the Demand Data Read requests sent to= uncore. Use it in conjunction with OFFCORE_REQUESTS_OUTSTANDING to determi= ne average latency in the uncore.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -1109,7 +1103,7 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.DEMAND_RFO", - "PublicDescription": "Counts the demand RFO (read for ownership) r= equests including regular RFOs, locks, ItoM. Available PDIST counters: 0", + "PublicDescription": "Counts the demand RFO (read for ownership) r= equests including regular RFOs, locks, ItoM.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -1119,7 +1113,6 @@ "Deprecated": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD", - "PublicDescription": "This event is deprecated. Refer to new event= OFFCORE_REQUESTS_OUTSTANDING.DATA_RD Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -1129,7 +1122,6 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD", - "PublicDescription": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DAT= A_RD Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -1139,7 +1131,7 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_CODE= _RD", - "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1149,7 +1141,6 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA= _RD", - "PublicDescription": "Cycles where at least 1 outstanding demand d= ata read request is pending. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1159,7 +1150,6 @@ "CounterMask": "1", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO", - "PublicDescription": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEM= AND_RFO Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -1168,7 +1158,6 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DATA_RD", - "PublicDescription": "OFFCORE_REQUESTS_OUTSTANDING.DATA_RD Availab= le PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -1177,7 +1166,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_CODE_RD", - "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS. Available PDIST counters: 0", + "PublicDescription": "Counts the number of offcore outstanding Cod= e Reads transactions in the super queue every cycle. The 'Offcore outstandi= ng' state of the transaction lasts from the L2 miss until the sending trans= action completion to requestor (SQ deallocation). See the corresponding Uma= sk under OFFCORE_REQUESTS.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1186,7 +1175,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD", - "PublicDescription": "For every cycle, increments by the number of= outstanding demand data read requests pending. Requests are considered o= utstanding from the time they miss the core's L2 cache until the transactio= n completion message is sent to the requestor. Available PDIST counters: 0", + "PublicDescription": "For every cycle, increments by the number of= outstanding demand data read requests pending. Requests are considered o= utstanding from the time they miss the core's L2 cache until the transactio= n completion message is sent to the requestor.", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -1195,7 +1184,7 @@ "Counter": "0,1,2,3", "EventCode": "0x2c", "EventName": "SQ_MISC.BUS_LOCK", - "PublicDescription": "Counts the more expensive bus lock needed to= enforce cache coherency for certain memory accesses that need to be done a= tomically. Can be created by issuing an atomic instruction (via the LOCK p= refix) which causes a cache line split or accesses uncacheable memory. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts the more expensive bus lock needed to= enforce cache coherency for certain memory accesses that need to be done a= tomically. Can be created by issuing an atomic instruction (via the LOCK p= refix) which causes a cache line split or accesses uncacheable memory.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -1204,7 +1193,6 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.ANY", - "PublicDescription": "Counts the number of PREFETCHNTA, PREFETCHW,= PREFETCHT0, PREFETCHT1 or PREFETCHT2 instructions executed. Available PDIS= T counters: 0", "SampleAfterValue": "100003", "UMask": "0xf" }, @@ -1213,7 +1201,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.NTA", - "PublicDescription": "Counts the number of PREFETCHNTA instruction= s executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHNTA instruction= s executed.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -1222,7 +1210,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.PREFETCHW", - "PublicDescription": "Counts the number of PREFETCHW instructions = executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHW instructions = executed.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -1231,7 +1219,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.T0", - "PublicDescription": "Counts the number of PREFETCHT0 instructions= executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHT0 instructions= executed.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -1240,7 +1228,7 @@ "Counter": "0,1,2,3", "EventCode": "0x40", "EventName": "SW_PREFETCH_ACCESS.T1_T2", - "PublicDescription": "Counts the number of PREFETCHT1 or PREFETCHT= 2 instructions executed. Available PDIST counters: 0", + "PublicDescription": "Counts the number of PREFETCHT1 or PREFETCHT= 2 instructions executed.", "SampleAfterValue": "100003", "UMask": "0x4" } diff --git a/tools/perf/pmu-events/arch/x86/sapphirerapids/floating-point.j= son b/tools/perf/pmu-events/arch/x86/sapphirerapids/floating-point.json index 8c9207750c82..bc475e163227 100644 --- a/tools/perf/pmu-events/arch/x86/sapphirerapids/floating-point.json +++ b/tools/perf/pmu-events/arch/x86/sapphirerapids/floating-point.json @@ -5,7 +5,6 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.FPDIV_ACTIVE", - "PublicDescription": "ARITH.FPDIV_ACTIVE Available PDIST counters:= 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -14,7 +13,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.FP", - "PublicDescription": "Counts all microcode Floating Point assists.= Available PDIST counters: 0", + "PublicDescription": "Counts all microcode Floating Point assists.= ", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -23,7 +22,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.SSE_AVX_MIX", - "PublicDescription": "ASSISTS.SSE_AVX_MIX Available PDIST counters= : 0", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -32,7 +30,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_0", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_0 [This event is al= ias to FP_ARITH_DISPATCHED.V0] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -41,7 +38,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_1", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_1 [This event is al= ias to FP_ARITH_DISPATCHED.V1] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -50,7 +46,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.PORT_5", - "PublicDescription": "FP_ARITH_DISPATCHED.PORT_5 [This event is al= ias to FP_ARITH_DISPATCHED.V2] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -59,7 +54,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V0", - "PublicDescription": "FP_ARITH_DISPATCHED.V0 [This event is alias = to FP_ARITH_DISPATCHED.PORT_0] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -68,7 +62,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V1", - "PublicDescription": "FP_ARITH_DISPATCHED.V1 [This event is alias = to FP_ARITH_DISPATCHED.PORT_1] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -77,7 +70,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb3", "EventName": "FP_ARITH_DISPATCHED.V2", - "PublicDescription": "FP_ARITH_DISPATCHED.V2 [This event is alias = to FP_ARITH_DISPATCHED.PORT_5] Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -86,7 +78,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 2 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as the= y perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR re= gister need to be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 2 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as the= y perform 2 calculations per element. The DAZ and FTZ flags in the MXCSR re= gister need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -95,7 +87,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events. Available PDIST co= unters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -104,7 +96,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 = calculations per element. The DAZ and FTZ flags in the MXCSR register need = to be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 4 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 = calculations per element. The DAZ and FTZ flags in the MXCSR register need = to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -113,7 +105,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events. Available PDIST co= unters: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed single precis= ion floating-point instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX S= QRT RSQRT RCP DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count tw= ice as they perform 2 calculations per element. The DAZ and FTZ flags in th= e MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -122,7 +114,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.4_FLOPS", - "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision and 256-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 2 or/and 4 computation operations, one for each element. = Applies to SSE* and AVX* packed single precision floating-point and packed= double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL= DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB ins= tructions count twice as they perform 2 calculations per element. The DAZ a= nd FTZ flags in the MXCSR register need to be set when using these events. = Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 128-bit pack= ed single precision and 256-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 2 or/and 4 computation operations, one for each element. = Applies to SSE* and AVX* packed single precision floating-point and packed= double precision floating-point instructions: ADD SUB HADD HSUB SUBADD MUL= DIV MIN MAX RCP14 RSQRT14 SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB ins= tructions count twice as they perform 2 calculations per element. The DAZ a= nd FTZ flags in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x18" }, @@ -131,7 +123,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP14= FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calc= ulations per element. The DAZ and FTZ flags in the MXCSR register need to b= e set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed double precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 8 computation opera= tions, one for each element. Applies to SSE* and AVX* packed double precis= ion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP14= FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 calc= ulations per element. The DAZ and FTZ flags in the MXCSR register need to b= e set when using these events.", "SampleAfterValue": "100003", "UMask": "0x40" }, @@ -140,7 +132,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE", - "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 16 computation oper= ations, one for each element. Applies to SSE* and AVX* packed single preci= sion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP1= 4 FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 cal= culations per element. The DAZ and FTZ flags in the MXCSR register need to = be set when using these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational 512-bit pack= ed single precision floating-point instructions retired; some instructions = will count twice as noted below. Each count represents 16 computation oper= ations, one for each element. Applies to SSE* and AVX* packed single preci= sion floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT14 RCP1= 4 FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform 2 cal= culations per element. The DAZ and FTZ flags in the MXCSR register need to = be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -149,7 +141,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.8_FLOPS", - "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision and 512-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 8 computation operations, one for each element. Applies = to SSE* and AVX* packed single precision and double precision floating-poin= t instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RSQRT14= RCP RCP14 DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice= as they perform 2 calculations per element. The DAZ and FTZ flags in the M= XCSR register need to be set when using these events. Available PDIST count= ers: 0", + "PublicDescription": "Number of SSE/AVX computational 256-bit pack= ed single precision and 512-bit packed double precision floating-point ins= tructions retired; some instructions will count twice as noted below. Each= count represents 8 computation operations, one for each element. Applies = to SSE* and AVX* packed single precision and double precision floating-poin= t instructions: ADD SUB HADD HSUB SUBADD MUL DIV MIN MAX SQRT RSQRT RSQRT14= RCP RCP14 DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice= as they perform 2 calculations per element. The DAZ and FTZ flags in the M= XCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x60" }, @@ -158,7 +150,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR", - "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision and double precision floating-point instructions retired; some = instructions will count twice as noted below. Each count represents 1 comp= utational operation. Applies to SSE* and AVX* scalar single precision float= ing-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB= . FM(N)ADD/SUB instructions count twice as they perform 2 calculations per= element. The DAZ and FTZ flags in the MXCSR register need to be set when u= sing these events. Available PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision and double precision floating-point instructions retired; some = instructions will count twice as noted below. Each count represents 1 comp= utational operation. Applies to SSE* and AVX* scalar single precision float= ing-point instructions: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB= . FM(N)ADD/SUB instructions count twice as they perform 2 calculations per= element. The DAZ and FTZ flags in the MXCSR register need to be set when u= sing these events.", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -167,7 +159,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE", - "PublicDescription": "Number of SSE/AVX computational scalar doubl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar double precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions co= unt twice as they perform 2 calculations per element. The DAZ and FTZ flags= in the MXCSR register need to be set when using these events. Available PD= IST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar doubl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar double precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions co= unt twice as they perform 2 calculations per element. The DAZ and FTZ flags= in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -176,7 +168,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE", - "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar single precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instr= uctions count twice as they perform 2 calculations per element. The DAZ and= FTZ flags in the MXCSR register need to be set when using these events. Av= ailable PDIST counters: 0", + "PublicDescription": "Number of SSE/AVX computational scalar singl= e precision floating-point instructions retired; some instructions will cou= nt twice as noted below. Each count represents 1 computational operation. = Applies to SSE* and AVX* scalar single precision floating-point instruction= s: ADD SUB MUL DIV MIN MAX SQRT RSQRT RCP FM(N)ADD/SUB. FM(N)ADD/SUB instr= uctions count twice as they perform 2 calculations per element. The DAZ and= FTZ flags in the MXCSR register need to be set when using these events.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -185,7 +177,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc7", "EventName": "FP_ARITH_INST_RETIRED.VECTOR", - "PublicDescription": "Number of any Vector retired FP arithmetic i= nstructions. The DAZ and FTZ flags in the MXCSR register need to be set wh= en using these events. Available PDIST counters: 0", + "PublicDescription": "Number of any Vector retired FP arithmetic i= nstructions. The DAZ and FTZ flags in the MXCSR register need to be set wh= en using these events.", "SampleAfterValue": "1000003", "UMask": "0xfc" }, @@ -194,7 +186,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.128B_PACKED_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.128B_PACKED_HALF Avai= lable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -203,7 +194,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.256B_PACKED_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.256B_PACKED_HALF Avai= lable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -212,7 +202,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.512B_PACKED_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.512B_PACKED_HALF Avai= lable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -221,7 +210,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.COMPLEX_SCALAR_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.COMPLEX_SCALAR_HALF A= vailable PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -230,7 +218,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.SCALAR", - "PublicDescription": "FP_ARITH_INST_RETIRED2.SCALAR Available PDIS= T counters: 0", + "PublicDescription": "FP_ARITH_INST_RETIRED2.SCALAR", "SampleAfterValue": "100003", "UMask": "0x3" }, @@ -239,7 +227,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.SCALAR_HALF", - "PublicDescription": "FP_ARITH_INST_RETIRED2.SCALAR_HALF Available= PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -248,7 +235,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcf", "EventName": "FP_ARITH_INST_RETIRED2.VECTOR", - "PublicDescription": "FP_ARITH_INST_RETIRED2.VECTOR Available PDIS= T counters: 0", + "PublicDescription": "FP_ARITH_INST_RETIRED2.VECTOR", "SampleAfterValue": "100003", "UMask": "0x1c" } diff --git a/tools/perf/pmu-events/arch/x86/sapphirerapids/frontend.json b/= tools/perf/pmu-events/arch/x86/sapphirerapids/frontend.json index 9fe9d62b867a..793c486ffabe 100644 --- a/tools/perf/pmu-events/arch/x86/sapphirerapids/frontend.json +++ b/tools/perf/pmu-events/arch/x86/sapphirerapids/frontend.json @@ -4,7 +4,7 @@ "Counter": "0,1,2,3", "EventCode": "0x60", "EventName": "BACLEARS.ANY", - "PublicDescription": "Number of times the front-end is resteered w= hen it finds a branch instruction in a fetch line. This is called Unknown B= ranch which occurs for the first time a branch instruction is fetched or wh= en the branch is not tracked by the BPU (Branch Prediction Unit) anymore. A= vailable PDIST counters: 0", + "PublicDescription": "Number of times the front-end is resteered w= hen it finds a branch instruction in a fetch line. This is called Unknown B= ranch which occurs for the first time a branch instruction is fetched or wh= en the branch is not tracked by the BPU (Branch Prediction Unit) anymore.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -13,7 +13,7 @@ "Counter": "0,1,2,3", "EventCode": "0x87", "EventName": "DECODE.LCP", - "PublicDescription": "Counts cycles that the Instruction Length de= coder (ILD) stalls occurred due to dynamically changing prefix length of th= e decoded instruction (by operand size prefix instruction 0x66, address siz= e prefix instruction 0x67 or REX.W for Intel64). Count is proportional to t= he number of prefixes in a 16B-line. This may result in a three-cycle penal= ty for each LCP (Length changing prefix) in a 16-byte chunk. Available PDIS= T counters: 0", + "PublicDescription": "Counts cycles that the Instruction Length de= coder (ILD) stalls occurred due to dynamically changing prefix length of th= e decoded instruction (by operand size prefix instruction 0x66, address siz= e prefix instruction 0x67 or REX.W for Intel64). Count is proportional to t= he number of prefixes in a 16B-line. This may result in a three-cycle penal= ty for each LCP (Length changing prefix) in a 16-byte chunk.", "SampleAfterValue": "500009", "UMask": "0x1" }, @@ -22,7 +22,6 @@ "Counter": "0,1,2,3", "EventCode": "0x87", "EventName": "DECODE.MS_BUSY", - "PublicDescription": "Cycles the Microcode Sequencer is busy. Avai= lable PDIST counters: 0", "SampleAfterValue": "500009", "UMask": "0x2" }, @@ -31,7 +30,7 @@ "Counter": "0,1,2,3", "EventCode": "0x61", "EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES", - "PublicDescription": "Decode Stream Buffer (DSB) is a Uop-cache th= at holds translations of previously fetched instructions that were decoded = by the legacy x86 decode pipeline (MITE). This event counts fetch penalty c= ycles when a transition occurs from DSB to MITE. Available PDIST counters: = 0", + "PublicDescription": "Decode Stream Buffer (DSB) is a Uop-cache th= at holds translations of previously fetched instructions that were decoded = by the legacy x86 decode pipeline (MITE). This event counts fetch penalty c= ycles when a transition occurs from DSB to MITE.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -249,7 +248,7 @@ "Counter": "0,1,2,3", "EventCode": "0x80", "EventName": "ICACHE_DATA.STALLS", - "PublicDescription": "Counts cycles where a code line fetch is sta= lled due to an L1 instruction cache miss. The decode pipeline works at a 32= Byte granularity. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where a code line fetch is sta= lled due to an L1 instruction cache miss. The decode pipeline works at a 32= Byte granularity.", "SampleAfterValue": "500009", "UMask": "0x4" }, @@ -260,7 +259,6 @@ "EdgeDetect": "1", "EventCode": "0x80", "EventName": "ICACHE_DATA.STALL_PERIODS", - "PublicDescription": "ICACHE_DATA.STALL_PERIODS Available PDIST co= unters: 0", "SampleAfterValue": "500009", "UMask": "0x4" }, @@ -269,7 +267,7 @@ "Counter": "0,1,2,3", "EventCode": "0x83", "EventName": "ICACHE_TAG.STALLS", - "PublicDescription": "Counts cycles where a code fetch is stalled = due to L1 instruction cache tag miss. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where a code fetch is stalled = due to L1 instruction cache tag miss.", "SampleAfterValue": "200003", "UMask": "0x4" }, @@ -279,7 +277,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES_ANY", - "PublicDescription": "Counts the number of cycles uops were delive= red to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) p= ath. Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles uops were delive= red to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) p= ath.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -289,7 +287,7 @@ "CounterMask": "6", "EventCode": "0x79", "EventName": "IDQ.DSB_CYCLES_OK", - "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the D= SB (Decode Stream Buffer) path. Count includes uops that may 'bypass' the I= DQ. Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the D= SB (Decode Stream Buffer) path. Count includes uops that may 'bypass' the I= DQ.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -298,7 +296,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.DSB_UOPS", - "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Availab= le PDIST counters: 0", + "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -308,7 +306,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES_ANY", - "PublicDescription": "Counts the number of cycles uops were delive= red to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipe= line) path. During these cycles uops are not being delivered from the Decod= e Stream Buffer (DSB). Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles uops were delive= red to the Instruction Decode Queue (IDQ) from the MITE (legacy decode pipe= line) path. During these cycles uops are not being delivered from the Decod= e Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -318,7 +316,7 @@ "CounterMask": "6", "EventCode": "0x79", "EventName": "IDQ.MITE_CYCLES_OK", - "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the M= ITE (legacy decode pipeline) path. During these cycles uops are not being d= elivered from the Decode Stream Buffer (DSB). Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles where optimal nu= mber of uops was delivered to the Instruction Decode Queue (IDQ) from the M= ITE (legacy decode pipeline) path. During these cycles uops are not being d= elivered from the Decode Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -327,7 +325,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MITE_UOPS", - "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the MITE path. This also means that uops are= not being delivered from the Decode Stream Buffer (DSB). Available PDIST c= ounters: 0", + "PublicDescription": "Counts the number of uops delivered to Instr= uction Decode Queue (IDQ) from the MITE path. This also means that uops are= not being delivered from the Decode Stream Buffer (DSB).", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -337,7 +335,7 @@ "CounterMask": "1", "EventCode": "0x79", "EventName": "IDQ.MS_CYCLES_ANY", - "PublicDescription": "Counts cycles during which uops are being de= livered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS= ) is busy. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts cycles during which uops are being de= livered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS= ) is busy. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.", "SampleAfterValue": "2000003", "UMask": "0x20" }, @@ -348,7 +346,7 @@ "EdgeDetect": "1", "EventCode": "0x79", "EventName": "IDQ.MS_SWITCHES", - "PublicDescription": "Number of switches from DSB (Decode Stream B= uffer) or MITE (legacy decode pipeline) to the Microcode Sequencer. Availab= le PDIST counters: 0", + "PublicDescription": "Number of switches from DSB (Decode Stream B= uffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -357,7 +355,7 @@ "Counter": "0,1,2,3", "EventCode": "0x79", "EventName": "IDQ.MS_UOPS", - "PublicDescription": "Counts the total number of uops delivered by= the Microcode Sequencer (MS). Available PDIST counters: 0", + "PublicDescription": "Counts the total number of uops delivered by= the Microcode Sequencer (MS).", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -366,7 +364,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CORE", - "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CORE] Available PDI= ST counters: 0", + "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -376,7 +374,7 @@ "CounterMask": "6", "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CYCLES_0_UOPS_DELIV.CORE", - "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CYCLES= _0_UOPS_DELIV.CORE] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_UOPS_NOT_DELIVERED.CYCLES= _0_UOPS_DELIV.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -387,7 +385,7 @@ "EventCode": "0x9c", "EventName": "IDQ_BUBBLES.CYCLES_FE_WAS_OK", "Invert": "1", - "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_UOPS_N= OT_DELIVERED.CYCLES_FE_WAS_OK] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_UOPS_N= OT_DELIVERED.CYCLES_FE_WAS_OK]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -396,7 +394,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CORE", - "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_BUBBLES.CORE] Available PDIST counters= : 0", + "PublicDescription": "Counts the number of uops not delivered to b= y the Instruction Decode Queue (IDQ) to the back-end of the pipeline when t= here was no back-end stalls. This event counts for one SMT thread in a give= n cycle. [This event is alias to IDQ_BUBBLES.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -406,7 +404,7 @@ "CounterMask": "6", "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE", - "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_BUBBLES.CYCLES_0_UOPS_DEL= IV.CORE] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when no uops wer= e delivered by the Instruction Decode Queue (IDQ) to the back-end of the pi= peline when there was no back-end stalls. This event counts for one SMT thr= ead in a given cycle. [This event is alias to IDQ_BUBBLES.CYCLES_0_UOPS_DEL= IV.CORE]", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -417,7 +415,7 @@ "EventCode": "0x9c", "EventName": "IDQ_UOPS_NOT_DELIVERED.CYCLES_FE_WAS_OK", "Invert": "1", - "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_BUBBLE= S.CYCLES_FE_WAS_OK] Available PDIST counters: 0", + "PublicDescription": "Counts the number of cycles when the optimal= number of uops were delivered by the Instruction Decode Queue (IDQ) to the= back-end of the pipeline when there was no back-end stalls. This event cou= nts for one SMT thread in a given cycle. [This event is alias to IDQ_BUBBLE= S.CYCLES_FE_WAS_OK]", "SampleAfterValue": "1000003", "UMask": "0x1" } diff --git a/tools/perf/pmu-events/arch/x86/sapphirerapids/memory.json b/to= ols/perf/pmu-events/arch/x86/sapphirerapids/memory.json index 7c3f9b76d367..5e6c1f05c981 100644 --- a/tools/perf/pmu-events/arch/x86/sapphirerapids/memory.json +++ b/tools/perf/pmu-events/arch/x86/sapphirerapids/memory.json @@ -5,7 +5,6 @@ "CounterMask": "6", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L3_MISS", - "PublicDescription": "Execution stalls while L3 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x6" }, @@ -14,7 +13,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.MEMORY_ORDERING", - "PublicDescription": "Counts the number of Machine Clears detected= dye to memory ordering. Memory Ordering Machine Clears may apply when a me= mory read may not conform to the memory ordering rules of the x86 architect= ure Available PDIST counters: 0", + "PublicDescription": "Counts the number of Machine Clears detected= dye to memory ordering. Memory Ordering Machine Clears may apply when a me= mory read may not conform to the memory ordering rules of the x86 architect= ure", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -24,7 +23,6 @@ "CounterMask": "2", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.CYCLES_L1D_MISS", - "PublicDescription": "Cycles while L1 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -34,7 +32,6 @@ "CounterMask": "3", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L1D_MISS", - "PublicDescription": "Execution stalls while L1 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -44,7 +41,7 @@ "CounterMask": "5", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L2_MISS", - "PublicDescription": "Execution stalls while L2 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock). Available PDIST counters: 0", + "PublicDescription": "Execution stalls while L2 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock).", "SampleAfterValue": "1000003", "UMask": "0x5" }, @@ -54,7 +51,7 @@ "CounterMask": "9", "EventCode": "0x47", "EventName": "MEMORY_ACTIVITY.STALLS_L3_MISS", - "PublicDescription": "Execution stalls while L3 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock). Available PDIST counters: 0", + "PublicDescription": "Execution stalls while L3 cache miss demand = cacheable load request is outstanding (will not count for uncacheable deman= d requests e.g. bus lock).", "SampleAfterValue": "1000003", "UMask": "0x9" }, @@ -478,7 +475,6 @@ "Counter": "0,1,2,3", "EventCode": "0x21", "EventName": "OFFCORE_REQUESTS.L3_MISS_DEMAND_DATA_RD", - "PublicDescription": "Counts demand data read requests that miss t= he L3 cache. Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -487,7 +483,7 @@ "Counter": "0,1,2,3", "EventCode": "0x20", "EventName": "OFFCORE_REQUESTS_OUTSTANDING.L3_MISS_DEMAND_DATA_RD", - "PublicDescription": "For every cycle, increments by the number of= demand data read requests pending that are known to have missed the L3 cac= he. Note that this does not capture all elapsed cycles while requests are = outstanding - only cycles from when the requests were known by the requesti= ng core to have missed the L3 cache. Available PDIST counters: 0", + "PublicDescription": "For every cycle, increments by the number of= demand data read requests pending that are known to have missed the L3 cac= he. Note that this does not capture all elapsed cycles while requests are = outstanding - only cycles from when the requests were known by the requesti= ng core to have missed the L3 cache.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -505,7 +501,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_EVENTS", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to none of the previous 3 categories (e.g. interrupt). Availabl= e PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to none of the previous 3 categories (e.g. interrupt).", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -514,7 +510,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MEM", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to various memory events (e.g. read/write capacity and conflict= s). Available PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to various memory events (e.g. read/write capacity and conflict= s).", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -523,7 +519,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_MEMTYPE", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to incompatible memory type. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to incompatible memory type.", "SampleAfterValue": "100003", "UMask": "0x40" }, @@ -532,7 +528,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.ABORTED_UNFRIENDLY", - "PublicDescription": "Counts the number of times an RTM execution = aborted due to HLE-unfriendly instructions. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times an RTM execution = aborted due to HLE-unfriendly instructions.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -541,7 +537,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.COMMIT", - "PublicDescription": "Counts the number of times RTM commit succee= ded. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times RTM commit succee= ded.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -550,7 +546,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc9", "EventName": "RTM_RETIRED.START", - "PublicDescription": "Counts the number of times we entered an RTM= region. Does not count nested transactions. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times we entered an RTM= region. Does not count nested transactions.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -559,7 +555,7 @@ "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CAPACITY_READ", - "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional reads Available PDIST counters: 0", + "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional reads", "SampleAfterValue": "100003", "UMask": "0x80" }, @@ -568,7 +564,7 @@ "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CAPACITY_WRITE", - "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional writes. Available PDIST counters: 0", + "PublicDescription": "Speculatively counts the number of Transacti= onal Synchronization Extensions (TSX) aborts due to a data capacity limitat= ion for transactional writes.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -577,7 +573,7 @@ "Counter": "0,1,2,3", "EventCode": "0x54", "EventName": "TX_MEM.ABORT_CONFLICT", - "PublicDescription": "Counts the number of times a TSX line had a = cache conflict. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times a TSX line had a = cache conflict.", "SampleAfterValue": "100003", "UMask": "0x1" } diff --git a/tools/perf/pmu-events/arch/x86/sapphirerapids/other.json b/too= ls/perf/pmu-events/arch/x86/sapphirerapids/other.json index a58d65556609..21f49f609ed4 100644 --- a/tools/perf/pmu-events/arch/x86/sapphirerapids/other.json +++ b/tools/perf/pmu-events/arch/x86/sapphirerapids/other.json @@ -4,10 +4,34 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.PAGE_FAULT", - "PublicDescription": "ASSISTS.PAGE_FAULT Available PDIST counters:= 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, + { + "BriefDescription": "HW_INTERRUPTS.MASKED", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xcb", + "EventName": "HW_INTERRUPTS.MASKED", + "SampleAfterValue": "100003", + "UMask": "0x2" + }, + { + "BriefDescription": "HW_INTERRUPTS.PENDING_AND_MASKED", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xcb", + "EventName": "HW_INTERRUPTS.PENDING_AND_MASKED", + "SampleAfterValue": "100003", + "UMask": "0x4" + }, + { + "BriefDescription": "Number of hardware interrupts received by the= processor.", + "Counter": "0,1,2,3,4,5,6,7", + "EventCode": "0xcb", + "EventName": "HW_INTERRUPTS.RECEIVED", + "PublicDescription": "Counts the number of hardware interruptions = received by the processor.", + "SampleAfterValue": "203", + "UMask": "0x1" + }, { "BriefDescription": "Counts streaming stores that have any type of= response.", "Counter": "0,1,2,3", @@ -25,7 +49,7 @@ "CounterMask": "1", "EventCode": "0x2d", "EventName": "XQ.FULL_CYCLES", - "PublicDescription": "number of cycles when the thread is active a= nd the uncore cannot take any further requests (for example prefetches, loa= ds or stores initiated by the Core that miss the L2 cache). Available PDIST= counters: 0", + "PublicDescription": "number of cycles when the thread is active a= nd the uncore cannot take any further requests (for example prefetches, loa= ds or stores initiated by the Core that miss the L2 cache).", "SampleAfterValue": "1000003", "UMask": "0x1" } diff --git a/tools/perf/pmu-events/arch/x86/sapphirerapids/pipeline.json b/= tools/perf/pmu-events/arch/x86/sapphirerapids/pipeline.json index 48bec483b49a..1fa7957956df 100644 --- a/tools/perf/pmu-events/arch/x86/sapphirerapids/pipeline.json +++ b/tools/perf/pmu-events/arch/x86/sapphirerapids/pipeline.json @@ -6,7 +6,6 @@ "Deprecated": "1", "EventCode": "0xb0", "EventName": "ARITH.DIVIDER_ACTIVE", - "PublicDescription": "This event is deprecated. Refer to new event= ARITH.DIV_ACTIVE Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x9" }, @@ -16,7 +15,7 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.DIV_ACTIVE", - "PublicDescription": "Counts cycles when divide unit is busy execu= ting divide or square root operations. Accounts for integer and floating-po= int operations. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when divide unit is busy execu= ting divide or square root operations. Accounts for integer and floating-po= int operations.", "SampleAfterValue": "1000003", "UMask": "0x9" }, @@ -27,7 +26,6 @@ "Deprecated": "1", "EventCode": "0xb0", "EventName": "ARITH.FP_DIVIDER_ACTIVE", - "PublicDescription": "This event is deprecated. Refer to new event= ARITH.FPDIV_ACTIVE Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -37,7 +35,6 @@ "CounterMask": "1", "EventCode": "0xb0", "EventName": "ARITH.IDIV_ACTIVE", - "PublicDescription": "This event counts the cycles the integer div= ider is busy. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -48,7 +45,6 @@ "Deprecated": "1", "EventCode": "0xb0", "EventName": "ARITH.INT_DIVIDER_ACTIVE", - "PublicDescription": "This event is deprecated. Refer to new event= ARITH.IDIV_ACTIVE Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -57,7 +53,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc1", "EventName": "ASSISTS.ANY", - "PublicDescription": "Counts the number of occurrences where a mic= rocode assist is invoked by hardware. Examples include AD (page Access Dirt= y), FP and AVX related assists. Available PDIST counters: 0", + "PublicDescription": "Counts the number of occurrences where a mic= rocode assist is invoked by hardware. Examples include AD (page Access Dirt= y), FP and AVX related assists.", "SampleAfterValue": "100003", "UMask": "0x1b" }, @@ -217,7 +213,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C01", - "PublicDescription": "Counts core clocks when the thread is in the= C0.1 light-weight slower wakeup time but more power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions. Availab= le PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.1 light-weight slower wakeup time but more power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -226,7 +222,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C02", - "PublicDescription": "Counts core clocks when the thread is in the= C0.2 light-weight faster wakeup time but less power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions. Availab= le PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.2 light-weight faster wakeup time but less power saving optimized state= . This state can be entered via the TPAUSE or UMWAIT instructions.", "SampleAfterValue": "2000003", "UMask": "0x20" }, @@ -235,7 +231,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.C0_WAIT", - "PublicDescription": "Counts core clocks when the thread is in the= C0.1 or C0.2 power saving optimized states (TPAUSE or UMWAIT instructions)= or running the PAUSE instruction. Available PDIST counters: 0", + "PublicDescription": "Counts core clocks when the thread is in the= C0.1 or C0.2 power saving optimized states (TPAUSE or UMWAIT instructions)= or running the PAUSE instruction.", "SampleAfterValue": "2000003", "UMask": "0x70" }, @@ -244,7 +240,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.DISTRIBUTED", - "PublicDescription": "This event distributes cycle counts between = active hyperthreads, i.e., those in C0. A hyperthread becomes inactive whe= n it executes the HLT or MWAIT instructions. If all other hyperthreads are= inactive (or disabled or do not exist), all counts are attributed to this = hyperthread. To obtain the full count when the Core is active, sum the coun= ts from each hyperthread. Available PDIST counters: 0", + "PublicDescription": "This event distributes cycle counts between = active hyperthreads, i.e., those in C0. A hyperthread becomes inactive whe= n it executes the HLT or MWAIT instructions. If all other hyperthreads are= inactive (or disabled or do not exist), all counts are attributed to this = hyperthread. To obtain the full count when the Core is active, sum the coun= ts from each hyperthread.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -253,7 +249,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE", - "PublicDescription": "Counts Core crystal clock cycles when curren= t thread is unhalted and the other thread is halted. Available PDIST counte= rs: 0", + "PublicDescription": "Counts Core crystal clock cycles when curren= t thread is unhalted and the other thread is halted.", "SampleAfterValue": "25003", "UMask": "0x2" }, @@ -262,7 +258,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.PAUSE", - "PublicDescription": "CPU_CLK_UNHALTED.PAUSE Available PDIST count= ers: 0", "SampleAfterValue": "2000003", "UMask": "0x40" }, @@ -273,7 +268,6 @@ "EdgeDetect": "1", "EventCode": "0xec", "EventName": "CPU_CLK_UNHALTED.PAUSE_INST", - "PublicDescription": "CPU_CLK_UNHALTED.PAUSE_INST Available PDIST = counters: 0", "SampleAfterValue": "2000003", "UMask": "0x40" }, @@ -282,7 +276,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.REF_DISTRIBUTED", - "PublicDescription": "This event distributes Core crystal clock cy= cle counts between active hyperthreads, i.e., those in C0 sleep-state. A hy= perthread becomes inactive when it executes the HLT or MWAIT instructions. = If one thread is active in a core, all counts are attributed to this hypert= hread. To obtain the full count when the Core is active, sum the counts fro= m each hyperthread. Available PDIST counters: 0", + "PublicDescription": "This event distributes Core crystal clock cy= cle counts between active hyperthreads, i.e., those in C0 sleep-state. A hy= perthread becomes inactive when it executes the HLT or MWAIT instructions. = If one thread is active in a core, all counts are attributed to this hypert= hread. To obtain the full count when the Core is active, sum the counts fro= m each hyperthread.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -299,7 +293,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.REF_TSC_P", - "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the four (eight when Hyperthr= eading is disabled) programmable counters available for other events. Note:= On all current platforms this event stops counting during 'throttling (TM)= ' states duty off periods the processor is 'halted'. The counter update is= done at a lower clock rate then the core clock the overflow status bit for= this counter may appear 'sticky'. After the counter has overflowed and so= ftware clears the overflow status bit and resets the counter to less than M= AX. The reset value to the counter is not clocked immediately so the overfl= ow status bit will flip 'high (1)' and generate another PMI (if enabled) af= ter which the reset value gets clocked into the counter. Therefore, softwar= e will get the interrupt, read the overflow status bit '1 for bit 34 while = the counter value is less than MAX. Software should ignore this case. Avail= able PDIST counters: 0", + "PublicDescription": "Counts the number of reference cycles when t= he core is not in a halt state. The core enters the halt state when it is r= unning the HLT instruction or the MWAIT instruction. This event is not affe= cted by core frequency changes (for example, P states, TM2 transitions) but= has the same incrementing frequency as the time stamp counter. This event = can approximate elapsed time while the core was not in a halt state. It is = counted on a dedicated fixed counter, leaving the four (eight when Hyperthr= eading is disabled) programmable counters available for other events. Note:= On all current platforms this event stops counting during 'throttling (TM)= ' states duty off periods the processor is 'halted'. The counter update is= done at a lower clock rate then the core clock the overflow status bit for= this counter may appear 'sticky'. After the counter has overflowed and so= ftware clears the overflow status bit and resets the counter to less than M= AX. The reset value to the counter is not clocked immediately so the overfl= ow status bit will flip 'high (1)' and generate another PMI (if enabled) af= ter which the reset value gets clocked into the counter. Therefore, softwar= e will get the interrupt, read the overflow status bit '1 for bit 34 while = the counter value is less than MAX. Software should ignore this case.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -316,7 +310,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0x3c", "EventName": "CPU_CLK_UNHALTED.THREAD_P", - "PublicDescription": "This is an architectural event that counts t= he number of thread cycles while the thread is not in a halt state. The thr= ead enters the halt state when it is running the HLT instruction. The core = frequency may change from time to time due to power or thermal throttling. = For this reason, this event may have a changing ratio with regards to wall = clock time. Available PDIST counters: 0", + "PublicDescription": "This is an architectural event that counts t= he number of thread cycles while the thread is not in a halt state. The thr= ead enters the halt state when it is running the HLT instruction. The core = frequency may change from time to time due to power or thermal throttling. = For this reason, this event may have a changing ratio with regards to wall = clock time.", "SampleAfterValue": "2000003" }, { @@ -325,7 +319,6 @@ "CounterMask": "8", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS", - "PublicDescription": "Cycles while L1 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x8" }, @@ -335,7 +328,6 @@ "CounterMask": "1", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS", - "PublicDescription": "Cycles while L2 cache miss demand load is ou= tstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -345,7 +337,6 @@ "CounterMask": "16", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY", - "PublicDescription": "Cycles while memory subsystem has an outstan= ding load. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -355,7 +346,6 @@ "CounterMask": "12", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS", - "PublicDescription": "Execution stalls while L1 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0xc" }, @@ -365,7 +355,6 @@ "CounterMask": "5", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS", - "PublicDescription": "Execution stalls while L2 cache miss demand = load is outstanding. Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x5" }, @@ -375,7 +364,6 @@ "CounterMask": "4", "EventCode": "0xa3", "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL", - "PublicDescription": "Total execution stalls. Available PDIST coun= ters: 0", "SampleAfterValue": "1000003", "UMask": "0x4" }, @@ -384,7 +372,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb7", "EventName": "EXE.AMX_BUSY", - "PublicDescription": "Counts the cycles where the AMX (Advance Mat= rix Extension) unit is busy performing an operation. Available PDIST counte= rs: 0", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -393,7 +380,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.1_PORTS_UTIL", - "PublicDescription": "Counts cycles during which a total of 1 uop = was executed on all ports and Reservation Station (RS) was not empty. Avail= able PDIST counters: 0", + "PublicDescription": "Counts cycles during which a total of 1 uop = was executed on all ports and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -402,7 +389,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.2_3_PORTS_UTIL", - "PublicDescription": "Cycles total of 2 or 3 uops are executed on = all ports and Reservation Station (RS) was not empty. Available PDIST count= ers: 0", "SampleAfterValue": "2000003", "UMask": "0xc" }, @@ -411,7 +397,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.2_PORTS_UTIL", - "PublicDescription": "Counts cycles during which a total of 2 uops= were executed on all ports and Reservation Station (RS) was not empty. Ava= ilable PDIST counters: 0", + "PublicDescription": "Counts cycles during which a total of 2 uops= were executed on all ports and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -420,7 +406,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.3_PORTS_UTIL", - "PublicDescription": "Cycles total of 3 uops are executed on all p= orts and Reservation Station (RS) was not empty. Available PDIST counters: = 0", + "PublicDescription": "Cycles total of 3 uops are executed on all p= orts and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -429,7 +415,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.4_PORTS_UTIL", - "PublicDescription": "Cycles total of 4 uops are executed on all p= orts and Reservation Station (RS) was not empty. Available PDIST counters: = 0", + "PublicDescription": "Cycles total of 4 uops are executed on all p= orts and Reservation Station (RS) was not empty.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -439,7 +425,6 @@ "CounterMask": "5", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.BOUND_ON_LOADS", - "PublicDescription": "Execution stalls while memory subsystem has = an outstanding load. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x21" }, @@ -449,7 +434,7 @@ "CounterMask": "2", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.BOUND_ON_STORES", - "PublicDescription": "Counts cycles where the Store Buffer was ful= l and no loads caused an execution stall. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where the Store Buffer was ful= l and no loads caused an execution stall.", "SampleAfterValue": "1000003", "UMask": "0x40" }, @@ -458,7 +443,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa6", "EventName": "EXE_ACTIVITY.EXE_BOUND_0_PORTS", - "PublicDescription": "Number of cycles total of 0 uops executed on= all ports, Reservation Station (RS) was not empty, the Store Buffer (SB) w= as not full and there was no outstanding load. Available PDIST counters: 0", + "PublicDescription": "Number of cycles total of 0 uops executed on= all ports, Reservation Station (RS) was not empty, the Store Buffer (SB) w= as not full and there was no outstanding load.", "SampleAfterValue": "1000003", "UMask": "0x80" }, @@ -467,7 +452,7 @@ "Counter": "0,1,2,3", "EventCode": "0x75", "EventName": "INST_DECODED.DECODERS", - "PublicDescription": "Number of decoders utilized in a cycle when = the MITE (legacy decode pipeline) fetches instructions. Available PDIST cou= nters: 0", + "PublicDescription": "Number of decoders utilized in a cycle when = the MITE (legacy decode pipeline) fetches instructions.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -492,7 +477,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.MACRO_FUSED", - "PublicDescription": "INST_RETIRED.MACRO_FUSED Available PDIST cou= nters: 0", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -501,7 +485,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.NOP", - "PublicDescription": "Counts all retired NOP or ENDBR32/64 instruc= tions Available PDIST counters: 0", + "PublicDescription": "Counts all retired NOP or ENDBR32/64 instruc= tions", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -518,7 +502,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc0", "EventName": "INST_RETIRED.REP_ITERATION", - "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent. Available PDIST counters: 0", + "PublicDescription": "Number of iterations of Repeat (REP) string = retired instructions such as MOVS, CMPS, and SCAS. Each has a byte, word, a= nd doubleword version and string instructions can be repeated using a repet= ition prefix, REP, that allows their architectural execution to be repeated= a number of times as specified by the RCX register. Note the number of ite= rations is implementation-dependent.", "SampleAfterValue": "2000003", "UMask": "0x8" }, @@ -529,7 +513,7 @@ "EdgeDetect": "1", "EventCode": "0xad", "EventName": "INT_MISC.CLEARS_COUNT", - "PublicDescription": "Counts the number of speculative clears due = to any type of branch misprediction or machine clears Available PDIST count= ers: 0", + "PublicDescription": "Counts the number of speculative clears due = to any type of branch misprediction or machine clears", "SampleAfterValue": "500009", "UMask": "0x1" }, @@ -538,7 +522,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.CLEAR_RESTEER_CYCLES", - "PublicDescription": "Cycles after recovery from a branch mispredi= ction or machine clear till the first uop is issued from the resteered path= . Available PDIST counters: 0", + "PublicDescription": "Cycles after recovery from a branch mispredi= ction or machine clear till the first uop is issued from the resteered path= .", "SampleAfterValue": "500009", "UMask": "0x80" }, @@ -547,7 +531,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.MBA_STALLS", - "PublicDescription": "INT_MISC.MBA_STALLS Available PDIST counters= : 0", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -556,7 +539,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.RECOVERY_CYCLES", - "PublicDescription": "Counts core cycles when the Resource allocat= or was stalled due to recovery from an earlier branch misprediction or mach= ine clear event. Available PDIST counters: 0", + "PublicDescription": "Counts core cycles when the Resource allocat= or was stalled due to recovery from an earlier branch misprediction or mach= ine clear event.", "SampleAfterValue": "500009", "UMask": "0x1" }, @@ -567,7 +550,6 @@ "EventName": "INT_MISC.UNKNOWN_BRANCH_CYCLES", "MSRIndex": "0x3F7", "MSRValue": "0x7", - "PublicDescription": "Bubble cycles of BAClear (Unknown Branch). A= vailable PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x40" }, @@ -576,7 +558,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xad", "EventName": "INT_MISC.UOP_DROPPING", - "PublicDescription": "Estimated number of Top-down Microarchitectu= re Analysis slots that got dropped due to non front-end reasons Available P= DIST counters: 0", + "PublicDescription": "Estimated number of Top-down Microarchitectu= re Analysis slots that got dropped due to non front-end reasons", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -585,7 +567,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.128BIT", - "PublicDescription": "INT_VEC_RETIRED.128BIT Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0x13" }, @@ -594,7 +575,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.256BIT", - "PublicDescription": "INT_VEC_RETIRED.256BIT Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0xac" }, @@ -603,7 +583,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.ADD_128", - "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 128-bit vector instructions. Available PDIST counters: 0= ", + "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 128-bit vector instructions.", "SampleAfterValue": "1000003", "UMask": "0x3" }, @@ -612,7 +592,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.ADD_256", - "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 256-bit vector instructions. Available PDIST counters: 0= ", + "PublicDescription": "Number of retired integer ADD/SUB (regular o= r horizontal), SAD 256-bit vector instructions.", "SampleAfterValue": "1000003", "UMask": "0xc" }, @@ -621,7 +601,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.MUL_256", - "PublicDescription": "INT_VEC_RETIRED.MUL_256 Available PDIST coun= ters: 0", "SampleAfterValue": "1000003", "UMask": "0x80" }, @@ -630,7 +609,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.SHUFFLES", - "PublicDescription": "INT_VEC_RETIRED.SHUFFLES Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x40" }, @@ -639,7 +617,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.VNNI_128", - "PublicDescription": "INT_VEC_RETIRED.VNNI_128 Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x10" }, @@ -648,7 +625,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe7", "EventName": "INT_VEC_RETIRED.VNNI_256", - "PublicDescription": "INT_VEC_RETIRED.VNNI_256 Available PDIST cou= nters: 0", "SampleAfterValue": "1000003", "UMask": "0x20" }, @@ -657,7 +633,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.ADDRESS_ALIAS", - "PublicDescription": "Counts the number of times a load got blocke= d due to false dependencies in MOB due to partial compare on address. Avail= able PDIST counters: 0", + "PublicDescription": "Counts the number of times a load got blocke= d due to false dependencies in MOB due to partial compare on address.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -666,7 +642,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.NO_SR", - "PublicDescription": "Counts the number of times that split load o= perations are temporarily blocked because all resources for handling the sp= lit accesses are in use. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times that split load o= perations are temporarily blocked because all resources for handling the sp= lit accesses are in use.", "SampleAfterValue": "100003", "UMask": "0x88" }, @@ -675,7 +651,7 @@ "Counter": "0,1,2,3", "EventCode": "0x03", "EventName": "LD_BLOCKS.STORE_FORWARD", - "PublicDescription": "Counts the number of times where store forwa= rding was prevented for a load operation. The most common case is a load bl= ocked due to the address of memory access (partially) overlapping with a pr= eceding uncompleted store. Note: See the table of not supported store forwa= rds in the Optimization Guide. Available PDIST counters: 0", + "PublicDescription": "Counts the number of times where store forwa= rding was prevented for a load operation. The most common case is a load bl= ocked due to the address of memory access (partially) overlapping with a pr= eceding uncompleted store. Note: See the table of not supported store forwa= rds in the Optimization Guide.", "SampleAfterValue": "100003", "UMask": "0x82" }, @@ -684,7 +660,7 @@ "Counter": "0,1,2,3", "EventCode": "0x4c", "EventName": "LOAD_HIT_PREFETCH.SWPF", - "PublicDescription": "Counts all software-prefetch load dispatches= that hit the fill buffer (FB) allocated for the software prefetch. It can = also be incremented by some lock instructions. So it should only be used wi= th profiling so that the locks can be excluded by ASM (Assembly File) inspe= ction of the nearby instructions. Available PDIST counters: 0", + "PublicDescription": "Counts all software-prefetch load dispatches= that hit the fill buffer (FB) allocated for the software prefetch. It can = also be incremented by some lock instructions. So it should only be used wi= th profiling so that the locks can be excluded by ASM (Assembly File) inspe= ction of the nearby instructions.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -694,7 +670,7 @@ "CounterMask": "1", "EventCode": "0xa8", "EventName": "LSD.CYCLES_ACTIVE", - "PublicDescription": "Counts the cycles when at least one uop is d= elivered by the LSD (Loop-stream detector). Available PDIST counters: 0", + "PublicDescription": "Counts the cycles when at least one uop is d= elivered by the LSD (Loop-stream detector).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -704,7 +680,7 @@ "CounterMask": "6", "EventCode": "0xa8", "EventName": "LSD.CYCLES_OK", - "PublicDescription": "Counts the cycles when optimal number of uop= s is delivered by the LSD (Loop-stream detector). Available PDIST counters:= 0", + "PublicDescription": "Counts the cycles when optimal number of uop= s is delivered by the LSD (Loop-stream detector).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -713,7 +689,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa8", "EventName": "LSD.UOPS", - "PublicDescription": "Counts the number of uops delivered to the b= ack-end by the LSD(Loop Stream Detector). Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops delivered to the b= ack-end by the LSD(Loop Stream Detector).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -724,7 +700,7 @@ "EdgeDetect": "1", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.COUNT", - "PublicDescription": "Counts the number of machine clears (nukes) = of any type. Available PDIST counters: 0", + "PublicDescription": "Counts the number of machine clears (nukes) = of any type.", "SampleAfterValue": "100003", "UMask": "0x1" }, @@ -733,7 +709,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc3", "EventName": "MACHINE_CLEARS.SMC", - "PublicDescription": "Counts self-modifying code (SMC) detected, w= hich causes a machine clear. Available PDIST counters: 0", + "PublicDescription": "Counts self-modifying code (SMC) detected, w= hich causes a machine clear.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -742,7 +718,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xe0", "EventName": "MISC2_RETIRED.LFENCE", - "PublicDescription": "number of LFENCE retired instructions Availa= ble PDIST counters: 0", + "PublicDescription": "number of LFENCE retired instructions", "SampleAfterValue": "400009", "UMask": "0x20" }, @@ -751,7 +727,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xcc", "EventName": "MISC_RETIRED.LBR_INSERTS", - "PublicDescription": "Increments when an entry is added to the Las= t Branch Record (LBR) array (or removed from the array in case of RETURNs i= n call stack mode). The event requires LBR enable via IA32_DEBUGCTL MSR and= branch type selection via MSR_LBR_SELECT. Available PDIST counters: 0", + "PublicDescription": "Increments when an entry is added to the Las= t Branch Record (LBR) array (or removed from the array in case of RETURNs i= n call stack mode). The event requires LBR enable via IA32_DEBUGCTL MSR and= branch type selection via MSR_LBR_SELECT.", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -760,7 +736,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa2", "EventName": "RESOURCE_STALLS.SB", - "PublicDescription": "Counts allocation stall cycles caused by the= store buffer (SB) being full. This counts cycles that the pipeline back-en= d blocked uop delivery from the front-end. Available PDIST counters: 0", + "PublicDescription": "Counts allocation stall cycles caused by the= store buffer (SB) being full. This counts cycles that the pipeline back-en= d blocked uop delivery from the front-end.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -769,7 +745,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa2", "EventName": "RESOURCE_STALLS.SCOREBOARD", - "PublicDescription": "Counts cycles where the pipeline is stalled = due to serializing operations. Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -778,7 +753,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa5", "EventName": "RS.EMPTY", - "PublicDescription": "Counts cycles during which the reservation s= tation (RS) is empty for this logical processor. This is usually caused whe= n the front-end pipeline runs into starvation periods (e.g. branch mispredi= ctions or i-cache misses) Available PDIST counters: 0", + "PublicDescription": "Counts cycles during which the reservation s= tation (RS) is empty for this logical processor. This is usually caused whe= n the front-end pipeline runs into starvation periods (e.g. branch mispredi= ctions or i-cache misses)", "SampleAfterValue": "1000003", "UMask": "0x7" }, @@ -790,7 +765,7 @@ "EventCode": "0xa5", "EventName": "RS.EMPTY_COUNT", "Invert": "1", - "PublicDescription": "Counts end of periods where the Reservation = Station (RS) was empty. Could be useful to closely sample on front-end late= ncy issues (see the FRONTEND_RETIRED event of designated precise events) Av= ailable PDIST counters: 0", + "PublicDescription": "Counts end of periods where the Reservation = Station (RS) was empty. Could be useful to closely sample on front-end late= ncy issues (see the FRONTEND_RETIRED event of designated precise events)", "SampleAfterValue": "100003", "UMask": "0x7" }, @@ -799,7 +774,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa5", "EventName": "RS.EMPTY_RESOURCE", - "PublicDescription": "Cycles when Reservation Station (RS) is empt= y due to a resource in the back-end Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -812,7 +786,6 @@ "EventCode": "0xa5", "EventName": "RS_EMPTY.COUNT", "Invert": "1", - "PublicDescription": "This event is deprecated. Refer to new event= RS.EMPTY_COUNT Available PDIST counters: 0", "SampleAfterValue": "100003", "UMask": "0x7" }, @@ -822,7 +795,6 @@ "Deprecated": "1", "EventCode": "0xa5", "EventName": "RS_EMPTY.CYCLES", - "PublicDescription": "This event is deprecated. Refer to new event= RS.EMPTY Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x7" }, @@ -831,7 +803,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.BACKEND_BOUND_SLOTS", - "PublicDescription": "Number of slots in TMA method where no micro= -operations were being issued from front-end to back-end of the machine due= to lack of back-end resources. Available PDIST counters: 0", + "PublicDescription": "Number of slots in TMA method where no micro= -operations were being issued from front-end to back-end of the machine due= to lack of back-end resources.", "SampleAfterValue": "10000003", "UMask": "0x2" }, @@ -840,7 +812,7 @@ "Counter": "0", "EventCode": "0xa4", "EventName": "TOPDOWN.BAD_SPEC_SLOTS", - "PublicDescription": "Number of slots of TMA method that were wast= ed due to incorrect speculation. It covers all types of control-flow or dat= a-related mis-speculations. Available PDIST counters: 0", + "PublicDescription": "Number of slots of TMA method that were wast= ed due to incorrect speculation. It covers all types of control-flow or dat= a-related mis-speculations.", "SampleAfterValue": "10000003", "UMask": "0x4" }, @@ -849,7 +821,7 @@ "Counter": "0", "EventCode": "0xa4", "EventName": "TOPDOWN.BR_MISPREDICT_SLOTS", - "PublicDescription": "Number of TMA slots that were wasted due to = incorrect speculation by (any type of) branch mispredictions. This event es= timates number of speculative operations that were issued but not retired a= s well as the out-of-order engine recovery past a branch misprediction. Ava= ilable PDIST counters: 0", + "PublicDescription": "Number of TMA slots that were wasted due to = incorrect speculation by (any type of) branch mispredictions. This event es= timates number of speculative operations that were issued but not retired a= s well as the out-of-order engine recovery past a branch misprediction.", "SampleAfterValue": "10000003", "UMask": "0x8" }, @@ -858,7 +830,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.MEMORY_BOUND_SLOTS", - "PublicDescription": "TOPDOWN.MEMORY_BOUND_SLOTS Available PDIST c= ounters: 0", "SampleAfterValue": "10000003", "UMask": "0x10" }, @@ -875,7 +846,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xa4", "EventName": "TOPDOWN.SLOTS_P", - "PublicDescription": "Counts the number of available slots for an = unhalted logical processor. The event increments by machine-width of the na= rrowest pipeline as employed by the Top-down Microarchitecture Analysis met= hod. The count is distributed among unhalted logical processors (hyper-thre= ads) who share the same physical core. Available PDIST counters: 0", + "PublicDescription": "Counts the number of available slots for an = unhalted logical processor. The event increments by machine-width of the na= rrowest pipeline as employed by the Top-down Microarchitecture Analysis met= hod. The count is distributed among unhalted logical processors (hyper-thre= ads) who share the same physical core.", "SampleAfterValue": "10000003", "UMask": "0x1" }, @@ -884,7 +855,6 @@ "Counter": "0,1,2,3", "EventCode": "0x76", "EventName": "UOPS_DECODED.DEC0_UOPS", - "PublicDescription": "UOPS_DECODED.DEC0_UOPS Available PDIST count= ers: 0", "SampleAfterValue": "1000003", "UMask": "0x1" }, @@ -893,7 +863,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_0", - "PublicDescription": "Number of uops dispatch to execution port 0= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 0= .", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -902,7 +872,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_1", - "PublicDescription": "Number of uops dispatch to execution port 1= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 1= .", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -911,7 +881,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_2_3_10", - "PublicDescription": "Number of uops dispatch to execution ports 2= , 3 and 10 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 2= , 3 and 10", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -920,7 +890,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_4_9", - "PublicDescription": "Number of uops dispatch to execution ports 4= and 9 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 4= and 9", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -929,7 +899,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_5_11", - "PublicDescription": "Number of uops dispatch to execution ports 5= and 11 Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports 5= and 11", "SampleAfterValue": "2000003", "UMask": "0x20" }, @@ -938,7 +908,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_6", - "PublicDescription": "Number of uops dispatch to execution port 6= . Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution port 6= .", "SampleAfterValue": "2000003", "UMask": "0x40" }, @@ -947,7 +917,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb2", "EventName": "UOPS_DISPATCHED.PORT_7_8", - "PublicDescription": "Number of uops dispatch to execution ports = 7 and 8. Available PDIST counters: 0", + "PublicDescription": "Number of uops dispatch to execution ports = 7 and 8.", "SampleAfterValue": "2000003", "UMask": "0x80" }, @@ -956,7 +926,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE", - "PublicDescription": "Counts the number of uops executed from any = thread. Available PDIST counters: 0", + "PublicDescription": "Counts the number of uops executed from any = thread.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -966,7 +936,7 @@ "CounterMask": "1", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1", - "PublicDescription": "Counts cycles when at least 1 micro-op is ex= ecuted from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 1 micro-op is ex= ecuted from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -976,7 +946,7 @@ "CounterMask": "2", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2", - "PublicDescription": "Counts cycles when at least 2 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 2 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -986,7 +956,7 @@ "CounterMask": "3", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3", - "PublicDescription": "Counts cycles when at least 3 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 3 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -996,7 +966,7 @@ "CounterMask": "4", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4", - "PublicDescription": "Counts cycles when at least 4 micro-ops are = executed from any thread on physical core. Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least 4 micro-ops are = executed from any thread on physical core.", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -1006,7 +976,7 @@ "CounterMask": "1", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_1", - "PublicDescription": "Cycles where at least 1 uop was executed per= -thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 1 uop was executed per= -thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1016,7 +986,7 @@ "CounterMask": "2", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_2", - "PublicDescription": "Cycles where at least 2 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 2 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1026,7 +996,7 @@ "CounterMask": "3", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_3", - "PublicDescription": "Cycles where at least 3 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 3 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1036,7 +1006,7 @@ "CounterMask": "4", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.CYCLES_GE_4", - "PublicDescription": "Cycles where at least 4 uops were executed p= er-thread. Available PDIST counters: 0", + "PublicDescription": "Cycles where at least 4 uops were executed p= er-thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1047,7 +1017,7 @@ "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.STALLS", "Invert": "1", - "PublicDescription": "Counts cycles during which no uops were disp= atched from the Reservation Station (RS) per thread. Available PDIST counte= rs: 0", + "PublicDescription": "Counts cycles during which no uops were disp= atched from the Reservation Station (RS) per thread.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1059,7 +1029,6 @@ "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.STALL_CYCLES", "Invert": "1", - "PublicDescription": "This event is deprecated. Refer to new event= UOPS_EXECUTED.STALLS Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1068,7 +1037,6 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.THREAD", - "PublicDescription": "Counts the number of uops to be executed per= -thread each cycle. Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1077,7 +1045,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xb1", "EventName": "UOPS_EXECUTED.X87", - "PublicDescription": "Counts the number of x87 uops executed. Avai= lable PDIST counters: 0", + "PublicDescription": "Counts the number of x87 uops executed.", "SampleAfterValue": "2000003", "UMask": "0x10" }, @@ -1086,7 +1054,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xae", "EventName": "UOPS_ISSUED.ANY", - "PublicDescription": "Counts the number of uops that the Resource = Allocation Table (RAT) issues to the Reservation Station (RS). Available PD= IST counters: 0", + "PublicDescription": "Counts the number of uops that the Resource = Allocation Table (RAT) issues to the Reservation Station (RS).", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1096,7 +1064,6 @@ "CounterMask": "1", "EventCode": "0xae", "EventName": "UOPS_ISSUED.CYCLES", - "PublicDescription": "UOPS_ISSUED.CYCLES Available PDIST counters:= 0", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1106,7 +1073,7 @@ "CounterMask": "1", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.CYCLES", - "PublicDescription": "Counts cycles where at least one uop has ret= ired. Available PDIST counters: 0", + "PublicDescription": "Counts cycles where at least one uop has ret= ired.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1115,7 +1082,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.HEAVY", - "PublicDescription": "Counts the number of retired micro-operation= s (uops) except the last uop of each instruction. An instruction that is de= coded into less than two uops does not contribute to the count. Available P= DIST counters: 0", + "PublicDescription": "Counts the number of retired micro-operation= s (uops) except the last uop of each instruction. An instruction that is de= coded into less than two uops does not contribute to the count.", "SampleAfterValue": "2000003", "UMask": "0x1" }, @@ -1126,7 +1093,6 @@ "EventName": "UOPS_RETIRED.MS", "MSRIndex": "0x3F7", "MSRValue": "0x8", - "PublicDescription": "UOPS_RETIRED.MS Available PDIST counters: 0", "SampleAfterValue": "2000003", "UMask": "0x4" }, @@ -1135,7 +1101,7 @@ "Counter": "0,1,2,3,4,5,6,7", "EventCode": "0xc2", "EventName": "UOPS_RETIRED.SLOTS", - "PublicDescription": "Counts the retirement slots used each cycle.= Available PDIST counters: 0", + "PublicDescription": "Counts the retirement slots used each cycle.= ", "SampleAfterValue": "2000003", "UMask": "0x2" }, @@ -1146,7 +1112,7 @@ "EventCode": "0xc2", "EventName": "UOPS_RETIRED.STALLS", "Invert": "1", - "PublicDescription": "This event counts cycles without actually re= tired uops. Available PDIST counters: 0", + "PublicDescription": "This event counts cycles without actually re= tired uops.", "SampleAfterValue": "1000003", "UMask": "0x2" }, @@ -1158,7 +1124,6 @@ "EventCode": "0xc2", "EventName": "UOPS_RETIRED.STALL_CYCLES", "Invert": "1", - "PublicDescription": "This event is deprecated. Refer to new event= UOPS_RETIRED.STALLS Available PDIST counters: 0", "SampleAfterValue": "1000003", "UMask": "0x2" } diff --git a/tools/perf/pmu-events/arch/x86/sapphirerapids/spr-metrics.json= b/tools/perf/pmu-events/arch/x86/sapphirerapids/spr-metrics.json index fe3f288be10e..8a18d6e01ef4 100644 --- a/tools/perf/pmu-events/arch/x86/sapphirerapids/spr-metrics.json +++ b/tools/perf/pmu-events/arch/x86/sapphirerapids/spr-metrics.json @@ -1,28 +1,28 @@ [ { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" @@ -90,6 +90,12 @@ "MetricName": "io_bandwidth_read", "ScaleUnit": "1MB/s" }, + { + "BriefDescription": "Bandwidth of inbound IO reads that are initia= ted by end device controllers that are requesting memory from the CPU and m= iss the L3 cache", + "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_MISS_PCIRDCUR * 64 / 1e6 / d= uration_time", + "MetricName": "io_bandwidth_read_l3_miss", + "ScaleUnit": "1MB/s" + }, { "BriefDescription": "Bandwidth of IO reads that are initiated by e= nd device controllers that are requesting memory from the local CPU socket", "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_PCIRDCUR_LOCAL * 64 / 1e6 / = duration_time", @@ -108,6 +114,12 @@ "MetricName": "io_bandwidth_write", "ScaleUnit": "1MB/s" }, + { + "BriefDescription": "Bandwidth of inbound IO writes that are initi= ated by end device controllers that are writing memory to the CPU", + "MetricExpr": "(UNC_CHA_TOR_INSERTS.IO_MISS_ITOM + UNC_CHA_TOR_INS= ERTS.IO_MISS_ITOMCACHENEAR) * 64 / 1e6 / duration_time", + "MetricName": "io_bandwidth_write_l3_miss", + "ScaleUnit": "1MB/s" + }, { "BriefDescription": "Bandwidth of IO writes that are initiated by = end device controllers that are writing memory to the local CPU socket", "MetricExpr": "(UNC_CHA_TOR_INSERTS.IO_ITOM_LOCAL + UNC_CHA_TOR_IN= SERTS.IO_ITOMCACHENEAR_LOCAL) * 64 / 1e6 / duration_time", @@ -123,19 +135,19 @@ { "BriefDescription": "Percentage of inbound full cacheline writes i= nitiated by end device controllers that miss the L3 cache", "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_MISS_ITOM / UNC_CHA_TOR_INSE= RTS.IO_ITOM", - "MetricName": "io_percent_of_inbound_full_writes_that_miss_l3", + "MetricName": "io_full_write_l3_miss", "ScaleUnit": "100%" }, { "BriefDescription": "Percentage of inbound partial cacheline write= s initiated by end device controllers that miss the L3 cache", "MetricExpr": "(UNC_CHA_TOR_INSERTS.IO_MISS_ITOMCACHENEAR + UNC_CH= A_TOR_INSERTS.IO_MISS_RFO) / (UNC_CHA_TOR_INSERTS.IO_ITOMCACHENEAR + UNC_CH= A_TOR_INSERTS.IO_RFO)", - "MetricName": "io_percent_of_inbound_partial_writes_that_miss_l3", + "MetricName": "io_partial_write_l3_miss", "ScaleUnit": "100%" }, { "BriefDescription": "Percentage of inbound reads initiated by end = device controllers that miss the L3 cache", "MetricExpr": "UNC_CHA_TOR_INSERTS.IO_MISS_PCIRDCUR / UNC_CHA_TOR_= INSERTS.IO_PCIRDCUR", - "MetricName": "io_percent_of_inbound_reads_that_miss_l3", + "MetricName": "io_read_l3_miss", "ScaleUnit": "100%" }, { @@ -395,7 +407,7 @@ { "BriefDescription": "This category represents fraction of slots wh= ere no uops are being delivered due to a lack of required resources for acc= epting new uops in the Backend", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "topdown\\-be\\-bound / (topdown\\-fe\\-bound + topd= own\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_inf= o_thread_slots", + "MetricExpr": "topdown\\-be\\-bound / (topdown\\-fe\\-bound + topd= own\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BvOB;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_backend_bound", "MetricThreshold": "tma_backend_bound > 0.2", @@ -429,40 +441,40 @@ "MetricThreshold": "tma_bottleneck_branching_overhead > 5", "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_amx_busy= + tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_c= ore_bound * tma_amx_busy / (tma_amx_busy + tma_divider + tma_ports_utilizat= ion + tma_serializing_operation) + tma_core_bound * (tma_ports_utilization = / (tma_amx_busy + tma_divider + tma_ports_utilization + tma_serializing_ope= ration)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_utili= zed_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_l= atency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)= ))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_late= ncy)) + tma_memory_bound * (tma_l3_bound / (tma_cxl_mem_bound + tma_dram_bo= und + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma= _sq_full / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_latency = + tma_sq_full)) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + t= ma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_boun= d)) * (tma_fb_full / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependen= cy + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependen= cy + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_= bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma= _l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_dtlb_load + tma_fb= _full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tm= a_store_fwd_blk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tm= a_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_split_l= oads / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_= latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_s= tore_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_sharin= g + tma_split_stores + tma_store_latency + tma_streaming_stores)) + tma_mem= ory_bound * (tma_store_bound / (tma_dram_bound + tma_l1_bound + tma_l2_boun= d + tma_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_dtlb_store= + tma_false_sharing + tma_split_stores + tma_store_latency + tma_streaming= _stores)))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latenc= y)) + tma_memory_bound * (tma_l3_bound / (tma_cxl_mem_bound + tma_dram_boun= d + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l= 3_hit_latency / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_lat= ency + tma_sq_full)) + tma_memory_bound * tma_l2_bound / (tma_cxl_mem_bound= + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_= bound) + tma_memory_bound * (tma_l1_bound / (tma_cxl_mem_bound + tma_dram_b= ound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tm= a_l1_latency_dependency / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dep= endency + tma_lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_me= mory_bound * (tma_l1_bound / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_b= ound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_lock_latency = / (tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_laten= cy + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_l1_bou= nd / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tm= a_l3_bound + tma_store_bound)) * (tma_split_loads / (tma_dtlb_load + tma_fb= _full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + tm= a_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_cxl_mem_boun= d + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store= _bound)) * (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tma_sp= lit_stores + tma_store_latency + tma_streaming_stores)) + tma_memory_bound = * (tma_store_bound / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + t= ma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_store_latency / (tma_= dtlb_store + tma_false_sharing + tma_split_stores + tma_store_latency + tma= _streaming_stores)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_amx_busy= + tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_c= ore_bound * tma_amx_busy / (tma_amx_busy + tma_divider + tma_ports_utilizat= ion + tma_serializing_operation) + tma_core_bound * (tma_ports_utilization = / (tma_amx_busy + tma_divider + tma_ports_utilization + tma_serializing_ope= ration)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_utili= zed_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - I= NST_RETIRED.REP_ITERATION / cpu@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetc= h_latency * (tma_ms_switches + tma_branch_resteers * (tma_clears_resteers += tma_mispredicts_resteers * tma_other_mispredicts / tma_branch_mispredicts)= / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_branches))= / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_m= isses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * tma_ms / (tma_ds= b + tma_mite + tma_ms))) - tma_bottleneck_big_code", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - (1 - I= NST_RETIRED.REP_ITERATION / cpu@UOPS_RETIRED.MS\\,cmask\\=3D1@) * (tma_fetc= h_latency * (tma_ms_switches + tma_branch_resteers * (tma_clears_resteers += tma_mispredicts_resteers * tma_other_mispredicts / tma_branch_mispredicts)= / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_branches))= / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_m= isses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottleneck_big_code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20" }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * ((1 - INST_RETIRED.REP_ITERATION / cpu@UOPS_R= ETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_switches + tma_bra= nch_resteers * (tma_clears_resteers + tma_mispredicts_resteers * tma_other_= mispredicts / tma_branch_mispredicts) / (tma_clears_resteers + tma_mispredi= cts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_dsb_swit= ches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) + t= ma_fetch_bandwidth * tma_ms / (tma_dsb + tma_mite + tma_ms)) + 10 * tma_mic= rocode_sequencer * tma_other_mispredicts / tma_branch_mispredicts * tma_bra= nch_mispredicts + tma_machine_clears * tma_other_nukes / tma_other_nukes + = tma_core_bound * (tma_serializing_operation + RS.EMPTY_RESOURCE / tma_info_= thread_clks * tma_ports_utilized_0) / (tma_amx_busy + tma_divider + tma_por= ts_utilization + tma_serializing_operation) + tma_microcode_sequencer / (tm= a_few_uops_instructions + tma_microcode_sequencer) * (tma_assists / tma_mic= rocode_sequencer) * tma_heavy_operations)", + "MetricExpr": "100 * ((1 - INST_RETIRED.REP_ITERATION / cpu@UOPS_R= ETIRED.MS\\,cmask\\=3D1@) * (tma_fetch_latency * (tma_ms_switches + tma_bra= nch_resteers * (tma_clears_resteers + tma_mispredicts_resteers * tma_other_= mispredicts / tma_branch_mispredicts) / (tma_clears_resteers + tma_mispredi= cts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_dsb_swit= ches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) + t= ma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_= mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_nukes= / tma_other_nukes + tma_core_bound * (tma_serializing_operation + RS.EMPTY= _RESOURCE / tma_info_thread_clks * tma_ports_utilized_0) / (tma_amx_busy + = tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_micr= ocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) * (= tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -470,7 +482,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", - "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + = tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, tma_dtlb_load + tma_= fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + = tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_dram_bound= + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_dt= lb_store / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_sto= re_latency + tma_streaming_stores)))", + "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bou= nd + tma_l3_bound + tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, = tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency = + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bou= nd / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tm= a_l3_bound + tma_store_bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_fa= lse_sharing + tma_split_stores + tma_store_latency + tma_streaming_stores))= )", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", "MetricThreshold": "tma_bottleneck_memory_data_tlbs > 20", @@ -478,7 +490,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Synchronization= related bottlenecks (data transfers and coherency updates across processor= s)", - "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * = (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) * tma_remote_cach= e / (tma_local_mem + tma_remote_cache + tma_remote_mem) + tma_l3_bound / (t= ma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_boun= d) * (tma_contested_accesses + tma_data_sharing) / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full) + tma_store_bound / = (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bo= und) * tma_false_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_= stores + tma_store_latency + tma_streaming_stores - tma_store_latency)) + t= ma_machine_clears * (1 - tma_other_nukes / tma_other_nukes))", + "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_cx= l_mem_bound + tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound) * (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency= )) * tma_remote_cache / (tma_local_mem + tma_remote_cache + tma_remote_mem)= + tma_l3_bound / (tma_cxl_mem_bound + tma_dram_bound + tma_l1_bound + tma_= l2_bound + tma_l3_bound + tma_store_bound) * (tma_contested_accesses + tma_= data_sharing) / (tma_contested_accesses + tma_data_sharing + tma_l3_hit_lat= ency + tma_sq_full) + tma_store_bound / (tma_cxl_mem_bound + tma_dram_bound= + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * tma_fals= e_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tma_st= ore_latency + tma_streaming_stores - tma_store_latency)) + tma_machine_clea= rs * (1 - tma_other_nukes / tma_other_nukes))", "MetricGroup": "BvMS;LockCont;Mem;Offcore;tma_issueSyncxn", "MetricName": "tma_bottleneck_memory_synchronization", "MetricThreshold": "tma_bottleneck_memory_synchronization > 10", @@ -494,7 +506,7 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -510,7 +522,7 @@ { "BriefDescription": "This metric represents fraction of slots the = CPU has wasted due to Branch Misprediction", "DefaultMetricgroupName": "TopdownL2", - "MetricExpr": "topdown\\-br\\-mispredict / (topdown\\-fe\\-bound += topdown\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tm= a_info_thread_slots", + "MetricExpr": "topdown\\-br\\-mispredict / (topdown\\-fe\\-bound += topdown\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BadSpec;BrMispredicts;BvMP;Default;TmaL2;TopdownL2= ;tma_L2_group;tma_bad_speculation_group;tma_issueBM", "MetricName": "tma_branch_mispredicts", "MetricThreshold": "tma_branch_mispredicts > 0.1 & tma_bad_specula= tion > 0.15", @@ -611,7 +623,6 @@ }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to contested acces= ses", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(76.6 * tma_info_system_core_frequency * (MEM_LOAD_= L3_HIT_RETIRED.XSNP_FWD * (OCR.DEMAND_DATA_RD.L3_HIT.SNOOP_HITM / (OCR.DEMA= ND_DATA_RD.L3_HIT.SNOOP_HITM + OCR.DEMAND_DATA_RD.L3_HIT.SNOOP_HIT_WITH_FWD= ))) + 74.6 * tma_info_system_core_frequency * MEM_LOAD_L3_HIT_RETIRED.XSNP_= MISS) * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS / 2) / tma_= info_thread_clks", "MetricGroup": "BvMS;DataSharing;LockCont;Offcore;Snoop;TopdownL4;= tma_L4_group;tma_issueSyncxn;tma_l3_bound_group", "MetricName": "tma_contested_accesses", @@ -630,6 +641,15 @@ "PublicDescription": "This metric represents fraction of slots whe= re Core non-memory issues were of a bottleneck. Shortage in hardware compu= te resources; or dependencies in software's instructions are both categoriz= ed under Core Bound. Hence it may indicate the machine ran out of an out-of= -order resource; certain execution units are overloaded or dependencies in = program's data- or instruction-flow are limiting the performance (e.g. FP-c= hained long-latency arithmetic operations).", "ScaleUnit": "100%" }, + { + "BriefDescription": "This metric roughly estimates (based on idle = latencies) how often the CPU was stalled on accesses to external CXL Memory= by loads (e.g", + "MetricExpr": "(((1 - ((19 * (MEM_LOAD_L3_MISS_RETIRED.REMOTE_DRAM= * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS)) + 10 * (MEM_LO= AD_L3_MISS_RETIRED.LOCAL_DRAM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RET= IRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REMOTE_FWD * (1 + MEM_LOAD_RETIRED= .FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REMOTE_HITM = * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS))) / (19 * (MEM_L= OAD_L3_MISS_RETIRED.REMOTE_DRAM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_R= ETIRED.L1_MISS)) + 10 * (MEM_LOAD_L3_MISS_RETIRED.LOCAL_DRAM * (1 + MEM_LOA= D_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LOAD_L3_MISS_RETIRED.REM= OTE_FWD * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS) + MEM_LO= AD_L3_MISS_RETIRED.REMOTE_HITM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RE= TIRED.L1_MISS)) + (25 * (MEM_LOAD_RETIRED.LOCAL_PMM * (1 + MEM_LOAD_RETIRED= .FB_HIT / MEM_LOAD_RETIRED.L1_MISS) if #has_pmem > 0 else 0) + 33 * (MEM_LO= AD_L3_MISS_RETIRED.REMOTE_PMM * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RET= IRED.L1_MISS) if #has_pmem > 0 else 0))) if #has_pmem > 0 else 1)) * (MEMOR= Y_ACTIVITY.STALLS_L3_MISS / tma_info_thread_clks) if 1e6 * (MEM_LOAD_L3_MIS= S_RETIRED.REMOTE_PMM + MEM_LOAD_RETIRED.LOCAL_PMM) > MEM_LOAD_RETIRED.L1_MI= SS else 0) if #has_pmem > 0 else 0)", + "MetricGroup": "MemoryBound;Server;TmaL3mem;TopdownL3;tma_L3_group= ;tma_memory_bound_group", + "MetricName": "tma_cxl_mem_bound", + "MetricThreshold": "tma_cxl_mem_bound > 0.1 & (tma_memory_bound > = 0.2 & tma_backend_bound > 0.2)", + "PublicDescription": "This metric roughly estimates (based on idle= latencies) how often the CPU was stalled on accesses to external CXL Memor= y by loads (e.g. 3D-Xpoint (Crystal Ridge, a.k.a. IXP) memory, PMM - Persis= tent Memory Module [from CLX to SPR] or any other CXL Type3 Memory [EMR onw= ards]).", + "ScaleUnit": "100%" + }, { "BriefDescription": "This metric estimates fraction of cycles whil= e the memory subsystem was handling synchronizations due to data-sharing ac= cesses", "MetricConstraint": "NO_GROUP_EVENTS", @@ -660,7 +680,7 @@ }, { "BriefDescription": "This metric estimates how often the CPU was s= talled on accesses to external memory (DRAM) by loads", - "MetricExpr": "MEMORY_ACTIVITY.STALLS_L3_MISS / tma_info_thread_cl= ks", + "MetricExpr": "(MEMORY_ACTIVITY.STALLS_L3_MISS / tma_info_thread_c= lks - tma_cxl_mem_bound if #has_pmem > 0 else MEMORY_ACTIVITY.STALLS_L3_MIS= S / tma_info_thread_clks)", "MetricGroup": "MemoryBound;TmaL3mem;TopdownL3;tma_L3_group;tma_me= mory_bound_group", "MetricName": "tma_dram_bound", "MetricThreshold": "tma_dram_bound > 0.1 & (tma_memory_bound > 0.2= & tma_backend_bound > 0.2)", @@ -718,7 +738,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%" }, { @@ -846,7 +866,7 @@ { "BriefDescription": "This metric represents fraction of slots wher= e the CPU was retiring heavy-weight operations -- instructions that require= two or more uops or micro-coded sequences", "DefaultMetricgroupName": "TopdownL2", - "MetricExpr": "topdown\\-heavy\\-ops / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_in= fo_thread_slots", + "MetricExpr": "topdown\\-heavy\\-ops / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "Default;Retire;TmaL2;TopdownL2;tma_L2_group;tma_re= tiring_group", "MetricName": "tma_heavy_operations", "MetricThreshold": "tma_heavy_operations > 0.1", @@ -1357,19 +1377,19 @@ { "BriefDescription": "Off-core accesses per kilo instruction for mo= dified write requests", "MetricExpr": "1e3 * OCR.MODIFIED_WRITE.ANY_RESPONSE / tma_info_in= st_mix_instructions", - "MetricGroup": "Offcore", + "MetricGroup": "Offcore;Server", "MetricName": "tma_info_memory_mix_offcore_mwrite_any_pki" }, { "BriefDescription": "Off-core accesses per kilo instruction for re= ads-to-core requests (speculative; including in-core HW prefetches)", "MetricExpr": "1e3 * OCR.READS_TO_CORE.ANY_RESPONSE / tma_info_ins= t_mix_instructions", - "MetricGroup": "CacheHits;Offcore", + "MetricGroup": "CacheHits;Offcore;Server", "MetricName": "tma_info_memory_mix_offcore_read_any_pki" }, { "BriefDescription": "L3 cache misses per kilo instruction for read= s-to-core requests (speculative; including in-core HW prefetches)", "MetricExpr": "1e3 * OCR.READS_TO_CORE.L3_MISS / tma_info_inst_mix= _instructions", - "MetricGroup": "Offcore", + "MetricGroup": "Offcore;Server", "MetricName": "tma_info_memory_mix_offcore_read_l3m_pki" }, { @@ -1395,21 +1415,21 @@ { "BriefDescription": "Average DRAM BW for Reads-to-Core (R2C) cover= ing for memory attached to local- and remote-socket", "MetricExpr": "64 * OCR.READS_TO_CORE.DRAM / 1e9 / tma_info_system= _time", - "MetricGroup": "HPC;Mem;MemoryBW;SoC", + "MetricGroup": "HPC;Mem;MemoryBW;Offcore;Server", "MetricName": "tma_info_memory_soc_r2c_dram_bw", "PublicDescription": "Average DRAM BW for Reads-to-Core (R2C) cove= ring for memory attached to local- and remote-socket. See R2C_Offcore_BW." }, { "BriefDescription": "Average L3-cache miss BW for Reads-to-Core (R= 2C)", "MetricExpr": "64 * OCR.READS_TO_CORE.L3_MISS / 1e9 / tma_info_sys= tem_time", - "MetricGroup": "HPC;Mem;MemoryBW;SoC", + "MetricGroup": "HPC;Mem;MemoryBW;Offcore;Server", "MetricName": "tma_info_memory_soc_r2c_l3m_bw", "PublicDescription": "Average L3-cache miss BW for Reads-to-Core (= R2C). This covering going to DRAM or other memory off-chip memory tears. Se= e R2C_Offcore_BW." }, { "BriefDescription": "Average Off-core access BW for Reads-to-Core = (R2C)", "MetricExpr": "64 * OCR.READS_TO_CORE.ANY_RESPONSE / 1e9 / tma_inf= o_system_time", - "MetricGroup": "HPC;Mem;MemoryBW;SoC", + "MetricGroup": "HPC;Mem;MemoryBW;Offcore;Server", "MetricName": "tma_info_memory_soc_r2c_offcore_bw", "PublicDescription": "Average Off-core access BW for Reads-to-Core= (R2C). R2C account for demand or prefetch load/RFO/code access that fill d= ata into the Core caches." }, @@ -1439,7 +1459,7 @@ "MetricName": "tma_info_memory_tlb_store_stlb_mpki" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (UOPS_EXECUTED.CORE_CYCLES_G= E_1 / 2 if #SMT_on else cpu@UOPS_EXECUTED.THREAD\\,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -1486,7 +1506,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -1498,16 +1518,28 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, + { + "BriefDescription": "Average 3DXP Memory Bandwidth Use for reads [= GB / sec]", + "MetricExpr": "(64 * UNC_M_PMM_RPQ_INSERTS / 1e9 / tma_info_system= _time if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryBW;Server;SoC", + "MetricName": "tma_info_system_cxl_mem_read_bw" + }, + { + "BriefDescription": "Average 3DXP Memory Bandwidth Use for Writes = [GB / sec]", + "MetricExpr": "(64 * UNC_M_PMM_WPQ_INSERTS / 1e9 / tma_info_system= _time if #has_pmem > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryBW;Server;SoC", + "MetricName": "tma_info_system_cxl_mem_write_bw" + }, { "BriefDescription": "Average external Memory Bandwidth Use for rea= ds and writes [GB / sec]", "MetricExpr": "64 * (UNC_M_CAS_COUNT.RD + UNC_M_CAS_COUNT.WR) / 1e= 9 / tma_info_system_time", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Giga Floating Point Operations Per Second", @@ -1571,9 +1603,15 @@ "MetricName": "tma_info_system_mem_parallel_reads", "PublicDescription": "Average number of parallel data read request= s to external memory. Accounts for demand loads and L1/L2 prefetches" }, + { + "BriefDescription": "Average latency of data read request to exter= nal 3D X-Point memory [in nanoseconds]", + "MetricExpr": "(1e9 * (UNC_CHA_TOR_OCCUPANCY.IA_MISS_DRD_PMM / UNC= _CHA_TOR_INSERTS.IA_MISS_DRD_PMM) / uncore_cha_0@event\\=3D0x1@ if #has_pme= m > 0 else 0)", + "MetricGroup": "MemOffcore;MemoryLat;Server;SoC", + "MetricName": "tma_info_system_mem_pmm_read_latency", + "PublicDescription": "Average latency of data read request to exte= rnal 3D X-Point memory [in nanoseconds]. Accounts for demand loads and L1/L= 2 data-read prefetches" + }, { "BriefDescription": "Average latency of data read request to exter= nal memory (in nanoseconds)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "1e9 * (UNC_CHA_TOR_OCCUPANCY.IA_MISS_DRD / UNC_CHA_= TOR_INSERTS.IA_MISS_DRD) / (tma_info_system_socket_clks / tma_info_system_t= ime)", "MetricGroup": "Mem;MemoryLat;SoC", "MetricName": "tma_info_system_mem_read_latency", @@ -1734,12 +1772,12 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", "MetricExpr": "min(2 * (MEM_INST_RETIRED.ALL_LOADS - MEM_LOAD_RETI= RED.FB_HIT - MEM_LOAD_RETIRED.L1_MISS) * 20 / 100, max(CYCLE_ACTIVITY.CYCLE= S_MEM_ANY - MEMORY_ACTIVITY.CYCLES_L1D_MISS, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%" }, { @@ -1753,7 +1791,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles wit= h demand load accesses that hit the L2 cache under unloaded scenarios (poss= ibly L2 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "4.4 * tma_info_system_core_frequency * MEM_LOAD_RET= IRED.L2_HIT * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS / 2) = / tma_info_thread_clks", "MetricGroup": "MemoryLat;TopdownL4;tma_L4_group;tma_l2_bound_grou= p", "MetricName": "tma_l2_hit_latency", @@ -1772,12 +1809,11 @@ }, { "BriefDescription": "This metric estimates fraction of cycles with= demand load accesses that hit the L3 cache under unloaded scenarios (possi= bly L3 latency limited)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "32.6 * tma_info_system_core_frequency * (MEM_LOAD_R= ETIRED.L3_HIT * (1 + MEM_LOAD_RETIRED.FB_HIT / MEM_LOAD_RETIRED.L1_MISS / 2= )) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%" }, { @@ -1860,6 +1896,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU spent handling cache misses due to lock operations", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(16 * max(0, MEM_INST_RETIRED.LOCK_LOADS - L2_RQSTS= .ALL_RFO) + MEM_INST_RETIRED.LOCK_LOADS / MEM_INST_RETIRED.ALL_STORES * (10= * L2_RQSTS.RFO_HIT + min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTAN= DING.CYCLES_WITH_DEMAND_RFO))) / tma_info_thread_clks", "MetricGroup": "LockCont;Offcore;TopdownL4;tma_L4_group;tma_issueR= FO;tma_l1_bound_group", "MetricName": "tma_lock_latency", @@ -1892,7 +1929,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%" }, { @@ -1901,13 +1938,13 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%" }, { "BriefDescription": "This metric represents fraction of slots the = Memory subsystem within the Backend was a bottleneck", "DefaultMetricgroupName": "TopdownL2", - "MetricExpr": "topdown\\-mem\\-bound / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_in= fo_thread_slots", + "MetricExpr": "topdown\\-mem\\-bound / (topdown\\-fe\\-bound + top= down\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "Backend;Default;TmaL2;TopdownL2;tma_L2_group;tma_b= ackend_bound_group", "MetricName": "tma_memory_bound", "MetricThreshold": "tma_memory_bound > 0.2 & tma_backend_bound > 0= .2", @@ -1917,7 +1954,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to LFENCE Instructions.", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * MISC2_RETIRED.LFENCE / tma_info_thread_clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_memory_fence", @@ -1970,7 +2006,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "max(IDQ.MS_CYCLES_ANY, cpu@UOPS_RETIRED.MS\\,cmask\= \=3D1@ / (UOPS_RETIRED.SLOTS / UOPS_ISSUED.ANY)) / tma_info_core_core_clks = / 2", + "MetricExpr": "max(IDQ.MS_CYCLES_ANY, cpu@UOPS_RETIRED.MS\\,cmask\= \=3D1@ / (UOPS_RETIRED.SLOTS / UOPS_ISSUED.ANY)) / tma_info_core_core_clks = / 2.4", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -2005,6 +2041,7 @@ }, { "BriefDescription": "This metric represents the remaining light uo= ps fraction the CPU has executed - remaining means not covered by other sib= ling nodes", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_light_operations - (tma_fp_arith + tma_i= nt_operations + tma_memory_operations + tma_fused_instructions + tma_non_fu= sed_branches))", "MetricGroup": "Pipeline;TopdownL3;tma_L3_group;tma_light_operatio= ns_group", "MetricName": "tma_other_light_ops", @@ -2066,6 +2103,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "((tma_ports_utilized_0 * tma_info_thread_clks + (EX= E_ACTIVITY.1_PORTS_UTIL + tma_retiring * EXE_ACTIVITY.2_3_PORTS_UTIL)) / tm= a_info_thread_clks if ARITH.DIV_ACTIVE < CYCLE_ACTIVITY.STALLS_TOTAL - EXE_= ACTIVITY.BOUND_ON_LOADS else (EXE_ACTIVITY.1_PORTS_UTIL + tma_retiring * EX= E_ACTIVITY.2_3_PORTS_UTIL) / tma_info_thread_clks)", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", @@ -2075,6 +2113,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed no uops on any execution port (Logical Processor cycles since ICL= , Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "(EXE_ACTIVITY.EXE_BOUND_0_PORTS + max(RS.EMPTY_RESO= URCE - RESOURCE_STALLS.SCOREBOARD, 0)) / tma_info_thread_clks * (CYCLE_ACTI= VITY.STALLS_TOTAL - EXE_ACTIVITY.BOUND_ON_LOADS) / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_ports_utiliza= tion_group", "MetricName": "tma_ports_utilized_0", @@ -2084,6 +2123,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles whe= re the CPU executed total of 1 uop per cycle on all execution ports (Logica= l Processor cycles since ICL, Physical Core cycles otherwise)", + "MetricConstraint": "NO_THRESHOLD_AND_NMI", "MetricExpr": "EXE_ACTIVITY.1_PORTS_UTIL / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issueL1;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_1", @@ -2093,7 +2133,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 2 uops per cycle on all execution ports (Logical Process= or cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "EXE_ACTIVITY.2_PORTS_UTIL / tma_info_thread_clks", "MetricGroup": "PortsUtil;TopdownL4;tma_L4_group;tma_issue2P;tma_p= orts_utilization_group", "MetricName": "tma_ports_utilized_2", @@ -2103,7 +2142,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles CPU= executed total of 3 or more uops per cycle on all execution ports (Logical= Processor cycles since ICL, Physical Core cycles otherwise)", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "UOPS_EXECUTED.CYCLES_GE_3 / tma_info_thread_clks", "MetricGroup": "BvCB;PortsUtil;TopdownL4;tma_L4_group;tma_ports_ut= ilization_group", "MetricName": "tma_ports_utilized_3m", @@ -2132,7 +2170,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_info_= thread_slots", + "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -2160,7 +2198,6 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to PAUSE Instructions", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "CPU_CLK_UNHALTED.PAUSE / tma_info_thread_clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_serializing_operation_g= roup", "MetricName": "tma_slow_pause", @@ -2192,7 +2229,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%" }, { diff --git a/tools/perf/pmu-events/arch/x86/sapphirerapids/uncore-memory.js= on b/tools/perf/pmu-events/arch/x86/sapphirerapids/uncore-memory.json index 68be01dad7c9..90f61c9511fc 100644 --- a/tools/perf/pmu-events/arch/x86/sapphirerapids/uncore-memory.json +++ b/tools/perf/pmu-events/arch/x86/sapphirerapids/uncore-memory.json @@ -2769,6 +2769,88 @@ "UMask": "0x3", "Unit": "iMC" }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.HIGH", + "Experimental": "1", + "PerPkg": "1", + "PublicDescription": "Number of DRAM Refreshes Issued : Counts the= number of refreshes issued.", + "UMask": "0x24", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.HIGH_ALL", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x24", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.HIGH_PCH0", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x4", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.HIGH_PCH1", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x20", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.PANIC", + "Experimental": "1", + "PerPkg": "1", + "PublicDescription": "Number of DRAM Refreshes Issued : Counts the= number of refreshes issued.", + "UMask": "0x12", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.PANIC_ALL", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x12", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.PANIC_PCH0", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x2", + "Unit": "iMC" + }, + { + "BriefDescription": "Number of DRAM Refreshes Issued", + "Counter": "0,1,2,3", + "EventCode": "0x45", + "EventName": "UNC_M_DRAM_REFRESH.PANIC_PCH1", + "Experimental": "1", + "PerPkg": "1", + "UMask": "0x10", + "Unit": "iMC" + }, { "BriefDescription": "ECC Correctable Errors", "Counter": "0,1,2,3", diff --git a/tools/perf/pmu-events/arch/x86/sapphirerapids/virtual-memory.j= son b/tools/perf/pmu-events/arch/x86/sapphirerapids/virtual-memory.json index 3d3f88600e26..609a9549cbf3 100644 --- a/tools/perf/pmu-events/arch/x86/sapphirerapids/virtual-memory.json +++ b/tools/perf/pmu-events/arch/x86/sapphirerapids/virtual-memory.json @@ -4,7 +4,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.STLB_HIT", - "PublicDescription": "Counts loads that miss the DTLB (Data TLB) a= nd hit the STLB (Second level TLB). Available PDIST counters: 0", + "PublicDescription": "Counts loads that miss the DTLB (Data TLB) a= nd hit the STLB (Second level TLB).", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -14,7 +14,7 @@ "CounterMask": "1", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a demand load. Available PDIST cou= nters: 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a demand load.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -23,7 +23,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data loads. This implies it missed in the DTLB and furth= er levels of TLB. The page walk can end with or without a fault. Available = PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data loads. This implies it missed in the DTLB and furth= er levels of TLB. The page walk can end with or without a fault.", "SampleAfterValue": "100003", "UMask": "0xe" }, @@ -32,7 +32,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_1G", - "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -41,7 +41,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data loads. This implies address translations missed in the= DTLB and further levels of TLB. The page walk can end with or without a fa= ult. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data loads. This implies address translations missed in the= DTLB and further levels of TLB. The page walk can end with or without a fa= ult.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -50,7 +50,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data loads. This implies address translations missed in the DT= LB and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -59,7 +59,7 @@ "Counter": "0,1,2,3", "EventCode": "0x12", "EventName": "DTLB_LOAD_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for a demand load in the PMH (Page Miss Handler) each cycle. Available PDIS= T counters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for a demand load in the PMH (Page Miss Handler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -68,7 +68,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.STLB_HIT", - "PublicDescription": "Counts stores that miss the DTLB (Data TLB) = and hit the STLB (2nd Level TLB). Available PDIST counters: 0", + "PublicDescription": "Counts stores that miss the DTLB (Data TLB) = and hit the STLB (2nd Level TLB).", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -78,7 +78,7 @@ "CounterMask": "1", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a store. Available PDIST counters:= 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a store.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -87,7 +87,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data stores. This implies it missed in the DTLB and furt= her levels of TLB. The page walk can end with or without a fault. Available= PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes= ) caused by demand data stores. This implies it missed in the DTLB and furt= her levels of TLB. The page walk can end with or without a fault.", "SampleAfterValue": "100003", "UMask": "0xe" }, @@ -96,7 +96,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_1G", - "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (1G sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t.", "SampleAfterValue": "100003", "UMask": "0x8" }, @@ -105,7 +105,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data stores. This implies address translations missed in th= e DTLB and further levels of TLB. The page walk can end with or without a f= ault. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M sizes) c= aused by demand data stores. This implies address translations missed in th= e DTLB and further levels of TLB. The page walk can end with or without a f= ault.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -114,7 +114,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K sizes) caus= ed by demand data stores. This implies address translations missed in the D= TLB and further levels of TLB. The page walk can end with or without a faul= t.", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -123,7 +123,7 @@ "Counter": "0,1,2,3", "EventCode": "0x13", "EventName": "DTLB_STORE_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for a store in the PMH (Page Miss Handler) each cycle. Available PDIST coun= ters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for a store in the PMH (Page Miss Handler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -132,7 +132,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.STLB_HIT", - "PublicDescription": "Counts instruction fetch requests that miss = the ITLB (Instruction TLB) and hit the STLB (Second-level TLB). Available P= DIST counters: 0", + "PublicDescription": "Counts instruction fetch requests that miss = the ITLB (Instruction TLB) and hit the STLB (Second-level TLB).", "SampleAfterValue": "100003", "UMask": "0x20" }, @@ -142,7 +142,7 @@ "CounterMask": "1", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_ACTIVE", - "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a code (instruction fetch) request= . Available PDIST counters: 0", + "PublicDescription": "Counts cycles when at least one PMH (Page Mi= ss Handler) is busy with a page walk for a code (instruction fetch) request= .", "SampleAfterValue": "100003", "UMask": "0x10" }, @@ -151,7 +151,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED", - "PublicDescription": "Counts completed page walks (all page sizes)= caused by a code fetch. This implies it missed in the ITLB (Instruction TL= B) and further levels of TLB. The page walk can end with or without a fault= . Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (all page sizes)= caused by a code fetch. This implies it missed in the ITLB (Instruction TL= B) and further levels of TLB. The page walk can end with or without a fault= .", "SampleAfterValue": "100003", "UMask": "0xe" }, @@ -160,7 +160,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED_2M_4M", - "PublicDescription": "Counts completed page walks (2M/4M page size= s) caused by a code fetch. This implies it missed in the ITLB (Instruction = TLB) and further levels of TLB. The page walk can end with or without a fau= lt. Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (2M/4M page size= s) caused by a code fetch. This implies it missed in the ITLB (Instruction = TLB) and further levels of TLB. The page walk can end with or without a fau= lt.", "SampleAfterValue": "100003", "UMask": "0x4" }, @@ -169,7 +169,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_COMPLETED_4K", - "PublicDescription": "Counts completed page walks (4K page sizes) = caused by a code fetch. This implies it missed in the ITLB (Instruction TLB= ) and further levels of TLB. The page walk can end with or without a fault.= Available PDIST counters: 0", + "PublicDescription": "Counts completed page walks (4K page sizes) = caused by a code fetch. This implies it missed in the ITLB (Instruction TLB= ) and further levels of TLB. The page walk can end with or without a fault.= ", "SampleAfterValue": "100003", "UMask": "0x2" }, @@ -178,7 +178,7 @@ "Counter": "0,1,2,3", "EventCode": "0x11", "EventName": "ITLB_MISSES.WALK_PENDING", - "PublicDescription": "Counts the number of page walks outstanding = for an outstanding code (instruction fetch) request in the PMH (Page Miss H= andler) each cycle. Available PDIST counters: 0", + "PublicDescription": "Counts the number of page walks outstanding = for an outstanding code (instruction fetch) request in the PMH (Page Miss H= andler) each cycle.", "SampleAfterValue": "100003", "UMask": "0x10" } --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pj1-f73.google.com (mail-pj1-f73.google.com [209.85.216.73]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id C574A3314D0 for ; Mon, 18 Aug 2025 19:06:20 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.216.73 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543983; cv=none; b=b+V6uOIQoO5JHENdH1D4VUcOtGd9Eq6msAkK74atPUEQ32/CF8ZiRN5dJJfxJ534LnMIQXSU5iJ2wHtOf/Cvvz1vXyff/OSKKUM3jkZjhpfhdEllEfnRwMNsOZwN4lkmCc5/qnJ34nOViyCyyWCOPclsEg9GFajmpq8r4rB+c68= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543983; c=relaxed/simple; bh=1jpuuTzjhqz3qDkfKHOu7ocIaPlgM25SmYekDft8Sjs=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=SFKO8iUdnRMz8K77teliAkdgiM8ERO5IDwNvSWU8MUO8PQWGehS6xPq5WfeDTGC5FP0HrX8c+rNR148WhTRfOUxF9TPFBNysRBSDDGqc7AQ7i7unHOKcoDw3nT0KUiHMJOscMw3gt/DpV7RngRDB0d0QIBjcl2uLQq9j8aJaH3U= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=irfdp+ep; arc=none smtp.client-ip=209.85.216.73 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="irfdp+ep" Received: by mail-pj1-f73.google.com with SMTP id 98e67ed59e1d1-323267b6c8eso10010254a91.1 for ; Mon, 18 Aug 2025 12:06:20 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543980; x=1756148780; darn=vger.kernel.org; h=to:from:subject:message-id:references:mime-version:in-reply-to:date :from:to:cc:subject:date:message-id:reply-to; bh=gRZrJ83jbIC3fyyYlD8BxhT6pZCcjqxNkMUQZE+n8nk=; b=irfdp+epT+yKHB6cEDmrbD9OqllmMJMIfQmoeFxLzX5ZSIVt4MqLls6lM45OqqAQPj whO28JGD3/jP1WhZWGkWkNClq2N/DCm/EnjuDgkm95QjoWFHJOeicjzzGqThmjaPAW32 EMicDUuni+GbBe9dMZc8lHoV5wxygPp8SIr9QAYDkkHgTPxwe99PGqHKdMPUcceO1FhZ v/z+NV7sNH7MU90QhaabPPW1MO5fuwPJwZ6ZUQPWmZ9EOhNpE6eKnAjpBHsUdTkWW8sc AbK81YOkkn2I0OJYnoXMuiLWj1MAn7RWh77OOmUCLoymHB6lo3MZ29iWwgEEm0ArTvR/ ZZfg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543980; x=1756148780; h=to:from:subject:message-id:references:mime-version:in-reply-to:date :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=gRZrJ83jbIC3fyyYlD8BxhT6pZCcjqxNkMUQZE+n8nk=; b=gt39KS87/4DcGQDLbZn2D3XyuIZ42oWU75QtzdaisrweldzCvOiki50NbJdqE3iyzL sRgMTAVfsIFpWESTo4vQcv9A2Yox7cZMth38x3dv77SrR+Adjf9I+ELNAFJK8fBjWagd WGwnrWhVakcQcH7oyv2VoEb3aI2dVAiqkWg/Ipk2s3LnunjtieJz8E9qZCTbKwYh4TLW 5iuXOsPAIkQi5vWVfYIY+yScPiiB0x1yo41EQu8cn27Y091hEh9OEm1thcApEF4MN0i9 3AaeRjElqYlUbYykWWW7uSvBReD1NK5ZpNlw6/0UFAxdx5pHKGQ8eLBF+CdYFUkHCxV0 A1jQ== X-Forwarded-Encrypted: i=1; AJvYcCVYSSPtag48uSafDLy8U1GMWRfu7t+kCrx+cjI3dzcXL75tefxU0rea9sWeUEWjK2EIvaRy22H+LSMAYLM=@vger.kernel.org X-Gm-Message-State: AOJu0YyH2Z2f3C4mRZdHUx3AlLwhkLc12vfMTLcbcLUBWyQf6bD4rGkv jUjr51i1bd2VYWY01n79OG75kIWsJvDLeiBCyVDBCYhXoqM1GnFMVG+ZNrDU4LukSPNDJvwNpi4 rwo9hf7yweA== X-Google-Smtp-Source: AGHT+IFe8/7n25l71Po2FXD0IXg8qcH8/D48vc/DuUu9lw5pVT8d2sJPZmLb/ZBnhjg6BkFhzwLKX9uN4M30 X-Received: from pjbta5.prod.google.com ([2002:a17:90b:4ec5:b0:2e0:915d:d594]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:90b:2e0f:b0:31f:42e8:a896 with SMTP id 98e67ed59e1d1-3237f60d711mr428628a91.34.1755543980004; Mon, 18 Aug 2025 12:06:20 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:14 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-19-irogers@google.com> Subject: [PATCH v2 18/20] perf vendor events: Update sierraforest metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/sierraforest/srf-metrics.json | 20 +++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/sierraforest/srf-metrics.json b= /tools/perf/pmu-events/arch/x86/sierraforest/srf-metrics.json index b9f3c611d87b..ca2c55917e55 100644 --- a/tools/perf/pmu-events/arch/x86/sierraforest/srf-metrics.json +++ b/tools/perf/pmu-events/arch/x86/sierraforest/srf-metrics.json @@ -1,56 +1,56 @@ [ { "BriefDescription": "C10 residency percent per package", - "MetricExpr": "cstate_pkg@c10\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c10\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C10_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C1 residency percent per core", - "MetricExpr": "cstate_core@c1\\-residency@ / TSC", + "MetricExpr": "cstate_core@c1\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C1_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C8 residency percent per package", - "MetricExpr": "cstate_pkg@c8\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c8\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C8_Pkg_Residency", "ScaleUnit": "100%" @@ -735,7 +735,7 @@ }, { "BriefDescription": "Average CPU Utilization", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricName": "tma_info_system_cpu_utilization" }, { @@ -747,7 +747,7 @@ }, { "BriefDescription": "Fraction of cycles spent in Kernel mode", - "MetricExpr": "cpu@CPU_CLK_UNHALTED.CORE_P@k / CPU_CLK_UNHALTED.CO= RE", + "MetricExpr": "CPU_CLK_UNHALTED.CORE_P:k / CPU_CLK_UNHALTED.CORE", "MetricGroup": "Summary", "MetricName": "tma_info_system_kernel_utilization" }, --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pg1-f201.google.com (mail-pg1-f201.google.com [209.85.215.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 303A9338F3E for ; Mon, 18 Aug 2025 19:06:26 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.215.201 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543991; cv=none; b=seE08IqhyCP2XSswtxooi3TlKICg2YtoqsImRlMoBrtZruiV6JGUdR/uMOXpY2TIWwS8Ki+0W8pS9/Jmtw7Fp5aDIWdbxaPmOJSacUh6ETv+dtjz5qVkagS7SdK8iKzmHSDf+vFVO7xOPw8BdvC5ToHgFA6QMUFgi1IQaUf0dcU= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543991; c=relaxed/simple; bh=c+vnH53dueGrXjkmBhCR4TCZuwVUSOh1cvKBTpQQNMs=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=qsuU3LPlS2e4vSUNQqLRURqZ+IKpTACURMYOEZB9JONMBF+EvFfx04y90sfuA3tzruCQirp3+OIDsGAxJMPfJEv4qtoyadw0U8GXaYuf11iMZSJMOcZka6A4ioyT/Vpr6gIvbP+XONnbxgnfbs0WjY/pfTghKM/7SbSG1ZQ4v+U= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=S2y3MI6X; arc=none smtp.client-ip=209.85.215.201 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="S2y3MI6X" Received: by mail-pg1-f201.google.com with SMTP id 41be03b00d2f7-b474d5ff588so722288a12.2 for ; Mon, 18 Aug 2025 12:06:26 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543985; x=1756148785; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=BzHAmaeXA9OIiXJ+xcmKvHI7wCRNulVu9l87+Mg8X58=; b=S2y3MI6XSU9a3xtUdgNKBDv/nOeIdWie15i+8BxPX+pja3+z9Q7mzn6+8nQlUnxTBE NcQOlwCEwYco3GGF69mrXGObvvL/Yvlh6/NGMi0WrwmJu2BfslDbmDEHmqjvzjwVLECp +/gZFlduVEHuqMR1d6qFXqMU1+jc4D6qetkUHWCNmth7l2tT9QqsyfGBjbNxh7alrMYf BGqfuMPfhb5SBVx4t3j6s7s5TFZeChwSkVx0OwdsSBDVLwMzRA3/1Ms/R3IMV+1WJ1+Q 98EZmaWUryRuegf7lye2r3mfdsxib8c8VIx55rkT7QfrP4cUpOLw6YqRmjPuJmY/PqWo 9idQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543985; x=1756148785; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=BzHAmaeXA9OIiXJ+xcmKvHI7wCRNulVu9l87+Mg8X58=; b=a1d5wvUL9OoUHX6Kj2zG4/l1sqB7reMI+3BzbdM7p144wi5OqV9RJIt433aA+8XCfH jw7Fx7ag2XhbKd5HKYFXWhOsqkRXi2O5YowO32tv+OUD8hpG7jfKGNAo4wPBM0tWTaCB Vzem0G+OZaMCng3wUQ4lRrcyTiXascSt6MaW873z+3p8LKe1eC6MAoMYaDpv76BDBNSl syQhc0GB+7eKZaygqLGHJzY+J+CJnDUw6eQomVX/YxGwUoRncO356SPO7dt1RCjQzX9K nFA0ZS1ne0yop+7fP47ShUnhbn0rQSnpHu5viJB4Bf0GZ3Att2/ENhkACNsz3wj88zL2 crnA== X-Forwarded-Encrypted: i=1; AJvYcCVIkrkkqHna7aHZLlMlBSpMQpVU18Ytb5B7YveVInkBruqVPcackjRwL5vhrmZk98gSemeZxZBWm8fFkUA=@vger.kernel.org X-Gm-Message-State: AOJu0YxHAKcPZG0HHO3yjSHExAyfxYVDYODSosZ6WygHyMXeFB3QSsqp EBVeMkQ2g8lcjlZLAuuU+Uuq7v6vbEUVBKBHrhitOl0BvXHaBxpvzpel9FgriZUeIi0hfBBSdMX ks7iYmvViSQ== X-Google-Smtp-Source: AGHT+IHVk3xwOEFPAJ+yrp6GaENL2yxmQp/jyPrGJOpv6fScF397QZbyMRL9ny28afa4CvnySYOzk02Re/7s X-Received: from plblh13.prod.google.com ([2002:a17:903:290d:b0:240:52d7:e88a]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:903:ac3:b0:242:9bbc:3647 with SMTP id d9443c01a7336-2449d001e87mr5052245ad.57.1755543985452; Mon, 18 Aug 2025 12:06:25 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:15 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-20-irogers@google.com> Subject: [PATCH v2 19/20] perf vendor events: Update skylake metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/skylake/skl-metrics.json | 101 ++++++++++++------ .../arch/x86/skylakex/skx-metrics.json | 101 ++++++++++++------ 2 files changed, 134 insertions(+), 68 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/skylake/skl-metrics.json b/tool= s/perf/pmu-events/arch/x86/skylake/skl-metrics.json index 2d3a037e88b5..707bd8790ed2 100644 --- a/tools/perf/pmu-events/arch/x86/skylake/skl-metrics.json +++ b/tools/perf/pmu-events/arch/x86/skylake/skl-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -80,6 +80,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_0 + UOPS_DISPATCHED_PORT= .PORT_1 + UOPS_DISPATCHED_PORT.PORT_5 + UOPS_DISPATCHED_PORT.PORT_6) / tma_= info_thread_slots", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_alu_op_utilization", @@ -117,6 +118,7 @@ }, { "BriefDescription": "Total pipeline cost of instruction fetch rela= ted bottlenecks by large code footprint programs (i-side cache; TLB and BTB= misses)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * tma_fetch_latency * (tma_itlb_misses + tma_ic= ache_misses + tma_unknown_branches) / (tma_branch_resteers + tma_dsb_switch= es + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches)", "MetricGroup": "BigFootprint;BvBC;Fed;Frontend;IcMiss;MemoryTLB", "MetricName": "tma_bottleneck_big_code", @@ -130,32 +132,36 @@ "MetricThreshold": "tma_bottleneck_branching_overhead > 5", "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_4k_aliasing + tma_dtlb_load + tma_= fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + = tma_store_fwd_blk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l= 1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_b= lk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + = tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_= 4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma= _lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * = (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_boun= d + tma_store_bound)) * (tma_split_loads / (tma_4k_aliasing + tma_dtlb_load= + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_l= oads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_dra= m_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * = (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency)) + tma_memory_bound * (tma_store_bound / (tma_dram_bou= nd + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_= store_latency / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tm= a_store_latency)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * tma_fetch_latency * (tma_ms_swit= ches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_resteer= s * (10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_misp= redicts)) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_b= ranches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + t= ma_itlb_misses + tma_lcp + tma_ms_switches)) - tma_bottleneck_big_code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", @@ -163,6 +169,7 @@ }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_clea= rs_resteers + tma_mispredicts_resteers * (10 * tma_microcode_sequencer * tm= a_other_mispredicts / tma_branch_mispredicts)) / (tma_clears_resteers + tma= _mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma= _dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_swit= ches) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_m= ispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_nukes = / tma_other_nukes + tma_core_bound * (tma_serializing_operation + tma_core_= bound * RS_EVENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_utilized_= 0) / (tma_divider + tma_ports_utilization + tma_serializing_operation) + tm= a_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequence= r) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", @@ -171,6 +178,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + = tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, tma_4k_aliasing + tm= a_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + = tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound= / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store= _bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_spli= t_stores + tma_store_latency)))", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", @@ -179,6 +187,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Synchronization= related bottlenecks (data transfers and coherency updates across processor= s)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_l3_bound / (tma_dram= _bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * (t= ma_contested_accesses + tma_data_sharing) / (tma_contested_accesses + tma_d= ata_sharing + tma_l3_hit_latency + tma_sq_full) + tma_store_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * = tma_false_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency - tma_store_latency)) + tma_machine_clears * (1 - tma_o= ther_nukes / tma_other_nukes))", "MetricGroup": "BvMS;LockCont;Mem;Offcore;tma_issueSyncxn", "MetricName": "tma_bottleneck_memory_synchronization", @@ -187,6 +196,7 @@ }, { "BriefDescription": "Total pipeline cost of Branch Misprediction r= elated bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (1 - 10 * tma_microcode_sequencer * tma_other= _mispredicts / tma_branch_mispredicts) * (tma_branch_mispredicts + tma_fetc= h_latency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switc= hes + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches))", "MetricGroup": "Bad;BadSpec;BrMispredicts;BvMP;tma_issueBM", "MetricName": "tma_bottleneck_mispredictions", @@ -195,7 +205,8 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -203,6 +214,7 @@ }, { "BriefDescription": "Total pipeline cost of \"useful operations\" = - the portion of Retiring category not covered by Branching_Overhead nor Ir= regular_Overhead.", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_retiring - (BR_INST_RETIRED.ALL_BRANCHES= + 2 * BR_INST_RETIRED.NEAR_CALL + INST_RETIRED.NOP) / tma_info_thread_slot= s - tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_se= quencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "BvUW;Ret", "MetricName": "tma_bottleneck_useful_work", @@ -230,6 +242,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU retired uops originated from CISC (complex instruction set computer) in= struction", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_microcode_sequencer - tma_assists)", "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_gro= up", "MetricName": "tma_cisc", @@ -391,7 +404,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%" }, { @@ -454,7 +467,6 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-= point (FP) vector uops fraction the CPU has retired aggregated across all v= ector widths", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "FP_ARITH_INST_RETIRED.VECTOR / UOPS_RETIRED.RETIRE_= SLOTS", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_= group;tma_issue2P", "MetricName": "tma_fp_vector", @@ -520,6 +532,7 @@ }, { "BriefDescription": "Branch Misprediction Cost: Cycles representin= g fraction of TMA slots wasted per non-speculative branch misprediction (re= tired JEClear)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "tma_bottleneck_mispredictions * tma_info_thread_slo= ts / 4 / BR_MISP_RETIRED.ALL_BRANCHES / 100", "MetricGroup": "Bad;BrMispredicts;tma_issueBM", "MetricName": "tma_info_bad_spec_branch_misprediction_cost", @@ -555,6 +568,7 @@ }, { "BriefDescription": "Total pipeline cost of DSB (uop cache) hits -= subset of the Instruction_Fetch_BW Bottleneck", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_frontend_bound * (tma_fetch_bandwidth / = (tma_fetch_bandwidth + tma_fetch_latency)) * (tma_dsb / (tma_dsb + tma_mite= )))", "MetricGroup": "DSB;Fed;FetchBW;tma_issueFB", "MetricName": "tma_info_botlnk_l2_dsb_bandwidth", @@ -572,6 +586,7 @@ }, { "BriefDescription": "Total pipeline cost of Instruction Cache miss= es - subset of the Big_Code Bottleneck", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_fetch_latency * tma_icache_misses / (tma= _branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_misses += tma_lcp + tma_ms_switches))", "MetricGroup": "Fed;FetchLat;IcMiss;tma_issueFL", "MetricName": "tma_info_botlnk_l2_ic_misses", @@ -713,7 +728,6 @@ }, { "BriefDescription": "Instructions per FP Arithmetic instruction (l= ower number means higher occurrence rate)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "INST_RETIRED.ANY / (FP_ARITH_INST_RETIRED.SCALAR + = FP_ARITH_INST_RETIRED.VECTOR)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_iparith", @@ -975,7 +989,7 @@ "MetricName": "tma_info_memory_tlb_store_stlb_mpki" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (UOPS_EXECUTED.CORE_CYCLES_G= E_1 / 2 if #SMT_on else cpu@UOPS_EXECUTED.THREAD\\,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -992,6 +1006,12 @@ "MetricGroup": "Fed;FetchBW", "MetricName": "tma_info_pipeline_fetch_mite" }, + { + "BriefDescription": "Average number of uops fetched from MS per cy= cle", + "MetricExpr": "IDQ.MS_UOPS / cpu@IDQ.MS_UOPS\\,cmask\\=3D1@", + "MetricGroup": "Fed;FetchLat;MicroSeq", + "MetricName": "tma_info_pipeline_fetch_ms" + }, { "BriefDescription": "Instructions per a microcode Assist invocatio= n", "MetricExpr": "INST_RETIRED.ANY / (FP_ASSIST.ANY + OTHER_ASSISTS.A= NY)", @@ -1008,7 +1028,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -1020,7 +1040,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -1029,7 +1049,7 @@ "MetricExpr": "64 * (UNC_ARB_TRK_REQUESTS.ALL + UNC_ARB_COH_TRK_RE= QUESTS.ALL) / 1e6 / tma_info_system_time / 1e3", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Giga Floating Point Operations Per Second", @@ -1175,12 +1195,13 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "min(2 * (MEM_INST_RETIRED.ALL_LOADS - MEM_LOAD_RETI= RED.FB_HIT - MEM_LOAD_RETIRED.L1_MISS) * 20 / 100, max(CYCLE_ACTIVITY.CYCLE= S_MEM_ANY - CYCLE_ACTIVITY.CYCLES_L1D_MISS, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%" }, { @@ -1217,7 +1238,7 @@ "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%" }, { @@ -1241,6 +1262,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution port for Load operations", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_2 + UOPS_DISPATCHED_PORT= .PORT_3 + UOPS_DISPATCHED_PORT.PORT_7 - UOPS_DISPATCHED_PORT.PORT_4) / (2 *= tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_load_op_utilization", @@ -1267,6 +1289,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 1 GB pages for= data load accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_load_stlb_miss * DTLB_LOAD_MISSES.WALK_COMPLETE= D_1G / (DTLB_LOAD_MISSES.WALK_COMPLETED_4K + DTLB_LOAD_MISSES.WALK_COMPLETE= D_2M_4M + DTLB_LOAD_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_load_stlb_mis= s_group", "MetricName": "tma_load_stlb_miss_1g", @@ -1275,6 +1298,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 2 or 4 MB page= s for data load accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_load_stlb_miss * DTLB_LOAD_MISSES.WALK_COMPLETE= D_2M_4M / (DTLB_LOAD_MISSES.WALK_COMPLETED_4K + DTLB_LOAD_MISSES.WALK_COMPL= ETED_2M_4M + DTLB_LOAD_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_load_stlb_mis= s_group", "MetricName": "tma_load_stlb_miss_2m", @@ -1283,6 +1307,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 4 KB pages for= data load accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_load_stlb_miss * DTLB_LOAD_MISSES.WALK_COMPLETE= D_4K / (DTLB_LOAD_MISSES.WALK_COMPLETED_4K + DTLB_LOAD_MISSES.WALK_COMPLETE= D_2M_4M + DTLB_LOAD_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_load_stlb_mis= s_group", "MetricName": "tma_load_stlb_miss_4k", @@ -1291,6 +1316,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU spent handling cache misses due to lock operations", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(12 * max(0, MEM_INST_RETIRED.LOCK_LOADS - L2_RQSTS= .ALL_RFO) + MEM_INST_RETIRED.LOCK_LOADS / MEM_INST_RETIRED.ALL_STORES * (9 = * L2_RQSTS.RFO_HIT + min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTAND= ING.CYCLES_WITH_DEMAND_RFO))) / tma_info_thread_clks", "MetricGroup": "LockCont;Offcore;TopdownL4;tma_L4_group;tma_issueR= FO;tma_l1_bound_group", "MetricName": "tma_lock_latency", @@ -1315,7 +1341,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%" }, { @@ -1324,7 +1350,7 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%" }, { @@ -1348,7 +1374,6 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU was retiring uops fetched by the Microcode Sequencer (MS) unit", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / UOPS_ISSUED.ANY * IDQ.M= S_UOPS / tma_info_thread_slots", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_heavy_operatio= ns_group;tma_issueMC;tma_issueMS", "MetricName": "tma_microcode_sequencer", @@ -1358,6 +1383,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to Branch Resteers as a result of Branch Misprediction= at execution stage", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES / (BR_MISP_RETIRED.ALL= _BRANCHES + MACHINE_CLEARS.COUNT) * INT_MISC.CLEAR_RESTEER_CYCLES / tma_inf= o_thread_clks", "MetricGroup": "BadSpec;BrMispredicts;BvMP;TopdownL4;tma_L4_group;= tma_branch_resteers_group;tma_issueBM", "MetricName": "tma_mispredicts_resteers", @@ -1412,6 +1438,7 @@ }, { "BriefDescription": "This metric represents the remaining light uo= ps fraction the CPU has executed - remaining means not covered by other sib= ling nodes", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_light_operations - (tma_fp_arith + tma_m= emory_operations + tma_fused_instructions + tma_non_fused_branches))", "MetricGroup": "Pipeline;TopdownL3;tma_L3_group;tma_light_operatio= ns_group", "MetricName": "tma_other_light_ops", @@ -1421,6 +1448,7 @@ }, { "BriefDescription": "This metric estimates fraction of slots the C= PU was stalled due to other cases of misprediction (non-retired x86 branche= s or other types).", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(tma_branch_mispredicts * (1 - BR_MISP_RETIRED.A= LL_BRANCHES / (INT_MISC.CLEARS_COUNT - MACHINE_CLEARS.COUNT)), 0.0001)", "MetricGroup": "BrMispredicts;BvIO;TopdownL3;tma_L3_group;tma_bran= ch_mispredicts_group", "MetricName": "tma_other_mispredicts", @@ -1429,6 +1457,7 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU has wasted due to Nukes (Machine Clears) not related to memory ordering= .", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(tma_machine_clears * (1 - MACHINE_CLEARS.MEMORY= _ORDERING / MACHINE_CLEARS.COUNT), 0.0001)", "MetricGroup": "BvIO;Machine_Clears;TopdownL3;tma_L3_group;tma_mac= hine_clears_group", "MetricName": "tma_other_nukes", @@ -1509,6 +1538,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "((tma_ports_utilized_0 * tma_info_thread_clks + (EX= E_ACTIVITY.1_PORTS_UTIL + tma_retiring * EXE_ACTIVITY.2_PORTS_UTIL)) / tma_= info_thread_clks if ARITH.DIVIDER_ACTIVE < CYCLE_ACTIVITY.STALLS_TOTAL - CY= CLE_ACTIVITY.STALLS_MEM_ANY else (EXE_ACTIVITY.1_PORTS_UTIL + tma_retiring = * EXE_ACTIVITY.2_PORTS_UTIL) / tma_info_thread_clks)", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", @@ -1595,7 +1625,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%" }, { @@ -1652,6 +1682,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 1 GB pages for= data store accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_store_stlb_miss * DTLB_STORE_MISSES.WALK_COMPLE= TED_1G / (DTLB_STORE_MISSES.WALK_COMPLETED_4K + DTLB_STORE_MISSES.WALK_COMP= LETED_2M_4M + DTLB_STORE_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_store_stlb_mi= ss_group", "MetricName": "tma_store_stlb_miss_1g", @@ -1660,6 +1691,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 2 or 4 MB page= s for data store accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_store_stlb_miss * DTLB_STORE_MISSES.WALK_COMPLE= TED_2M_4M / (DTLB_STORE_MISSES.WALK_COMPLETED_4K + DTLB_STORE_MISSES.WALK_C= OMPLETED_2M_4M + DTLB_STORE_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_store_stlb_mi= ss_group", "MetricName": "tma_store_stlb_miss_2m", @@ -1668,6 +1700,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 4 KB pages for= data store accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_store_stlb_miss * DTLB_STORE_MISSES.WALK_COMPLE= TED_4K / (DTLB_STORE_MISSES.WALK_COMPLETED_4K + DTLB_STORE_MISSES.WALK_COMP= LETED_2M_4M + DTLB_STORE_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_store_stlb_mi= ss_group", "MetricName": "tma_store_stlb_miss_4k", diff --git a/tools/perf/pmu-events/arch/x86/skylakex/skx-metrics.json b/too= ls/perf/pmu-events/arch/x86/skylakex/skx-metrics.json index 7cc7b076c3e2..56c2caaafef4 100644 --- a/tools/perf/pmu-events/arch/x86/skylakex/skx-metrics.json +++ b/tools/perf/pmu-events/arch/x86/skylakex/skx-metrics.json @@ -1,49 +1,49 @@ [ { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per core", - "MetricExpr": "cstate_core@c3\\-residency@ / TSC", + "MetricExpr": "cstate_core@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" @@ -301,6 +301,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution ports for ALU operations.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_0 + UOPS_DISPATCHED_PORT= .PORT_1 + UOPS_DISPATCHED_PORT.PORT_5 + UOPS_DISPATCHED_PORT.PORT_6) / tma_= info_thread_slots", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_alu_op_utilization", @@ -338,6 +339,7 @@ }, { "BriefDescription": "Total pipeline cost of instruction fetch rela= ted bottlenecks by large code footprint programs (i-side cache; TLB and BTB= misses)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * tma_fetch_latency * (tma_itlb_misses + tma_ic= ache_misses + tma_unknown_branches) / (tma_branch_resteers + tma_dsb_switch= es + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches)", "MetricGroup": "BigFootprint;BvBC;Fed;Frontend;IcMiss;MemoryTLB", "MetricName": "tma_bottleneck_big_code", @@ -351,32 +353,36 @@ "MetricThreshold": "tma_bottleneck_branching_overhead > 5", "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_4k_aliasing + tma_dtlb_load + tma_= fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + = tma_store_fwd_blk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l= 1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_b= lk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + = tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_= 4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma= _lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * = (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_boun= d + tma_store_bound)) * (tma_split_loads / (tma_4k_aliasing + tma_dtlb_load= + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_l= oads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_dra= m_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * = (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency)) + tma_memory_bound * (tma_store_bound / (tma_dram_bou= nd + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_= store_latency / (tma_dtlb_store + tma_false_sharing + tma_split_stores + tm= a_store_latency)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * tma_fetch_latency * (tma_ms_swit= ches + tma_branch_resteers * (tma_clears_resteers + tma_mispredicts_resteer= s * (10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_misp= redicts)) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_b= ranches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + t= ma_itlb_misses + tma_lcp + tma_ms_switches)) - tma_bottleneck_big_code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", @@ -384,6 +390,7 @@ }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_clea= rs_resteers + tma_mispredicts_resteers * (10 * tma_microcode_sequencer * tm= a_other_mispredicts / tma_branch_mispredicts)) / (tma_clears_resteers + tma= _mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma= _dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_swit= ches) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_m= ispredicts * tma_branch_mispredicts + tma_machine_clears * tma_other_nukes = / tma_other_nukes + tma_core_bound * (tma_serializing_operation + tma_core_= bound * RS_EVENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_utilized_= 0) / (tma_divider + tma_ports_utilization + tma_serializing_operation) + tm= a_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequence= r) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", @@ -392,6 +399,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + = tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, tma_4k_aliasing + tm= a_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + = tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound= / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store= _bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_spli= t_stores + tma_store_latency)))", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", @@ -400,6 +408,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Synchronization= related bottlenecks (data transfers and coherency updates across processor= s)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * = (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) * tma_remote_cach= e / (tma_local_mem + tma_remote_cache + tma_remote_mem) + tma_l3_bound / (t= ma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_boun= d) * (tma_contested_accesses + tma_data_sharing) / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full) + tma_store_bound / = (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bo= und) * tma_false_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_= stores + tma_store_latency - tma_store_latency)) + tma_machine_clears * (1 = - tma_other_nukes / tma_other_nukes))", "MetricGroup": "BvMS;LockCont;Mem;Offcore;tma_issueSyncxn", "MetricName": "tma_bottleneck_memory_synchronization", @@ -408,6 +417,7 @@ }, { "BriefDescription": "Total pipeline cost of Branch Misprediction r= elated bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (1 - 10 * tma_microcode_sequencer * tma_other= _mispredicts / tma_branch_mispredicts) * (tma_branch_mispredicts + tma_fetc= h_latency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switc= hes + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches))", "MetricGroup": "Bad;BadSpec;BrMispredicts;BvMP;tma_issueBM", "MetricName": "tma_bottleneck_mispredictions", @@ -416,7 +426,8 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -424,6 +435,7 @@ }, { "BriefDescription": "Total pipeline cost of \"useful operations\" = - the portion of Retiring category not covered by Branching_Overhead nor Ir= regular_Overhead.", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_retiring - (BR_INST_RETIRED.ALL_BRANCHES= + 2 * BR_INST_RETIRED.NEAR_CALL + INST_RETIRED.NOP) / tma_info_thread_slot= s - tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_se= quencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "BvUW;Ret", "MetricName": "tma_bottleneck_useful_work", @@ -451,6 +463,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU retired uops originated from CISC (complex instruction set computer) in= struction", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_microcode_sequencer - tma_assists)", "MetricGroup": "TopdownL4;tma_L4_group;tma_microcode_sequencer_gro= up", "MetricName": "tma_cisc", @@ -612,7 +625,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%" }, { @@ -675,7 +688,6 @@ }, { "BriefDescription": "This metric approximates arithmetic floating-= point (FP) vector uops fraction the CPU has retired aggregated across all v= ector widths", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umas= k\\=3D0xfc@ / UOPS_RETIRED.RETIRE_SLOTS", "MetricGroup": "Compute;Flops;TopdownL4;tma_L4_group;tma_fp_arith_= group;tma_issue2P", "MetricName": "tma_fp_vector", @@ -750,6 +762,7 @@ }, { "BriefDescription": "Branch Misprediction Cost: Cycles representin= g fraction of TMA slots wasted per non-speculative branch misprediction (re= tired JEClear)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "tma_bottleneck_mispredictions * tma_info_thread_slo= ts / 4 / BR_MISP_RETIRED.ALL_BRANCHES / 100", "MetricGroup": "Bad;BrMispredicts;tma_issueBM", "MetricName": "tma_info_bad_spec_branch_misprediction_cost", @@ -785,6 +798,7 @@ }, { "BriefDescription": "Total pipeline cost of DSB (uop cache) hits -= subset of the Instruction_Fetch_BW Bottleneck", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_frontend_bound * (tma_fetch_bandwidth / = (tma_fetch_bandwidth + tma_fetch_latency)) * (tma_dsb / (tma_dsb + tma_mite= )))", "MetricGroup": "DSB;Fed;FetchBW;tma_issueFB", "MetricName": "tma_info_botlnk_l2_dsb_bandwidth", @@ -802,6 +816,7 @@ }, { "BriefDescription": "Total pipeline cost of Instruction Cache miss= es - subset of the Big_Code Bottleneck", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_fetch_latency * tma_icache_misses / (tma= _branch_resteers + tma_dsb_switches + tma_icache_misses + tma_itlb_misses += tma_lcp + tma_ms_switches))", "MetricGroup": "Fed;FetchLat;IcMiss;tma_issueFL", "MetricName": "tma_info_botlnk_l2_ic_misses", @@ -943,7 +958,6 @@ }, { "BriefDescription": "Instructions per FP Arithmetic instruction (l= ower number means higher occurrence rate)", - "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "INST_RETIRED.ANY / (FP_ARITH_INST_RETIRED.SCALAR + = cpu@FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE\\,umask\\=3D0xfc@)", "MetricGroup": "Flops;InsType", "MetricName": "tma_info_inst_mix_iparith", @@ -1225,7 +1239,7 @@ "MetricName": "tma_info_memory_tlb_store_stlb_mpki" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (UOPS_EXECUTED.CORE_CYCLES_G= E_1 / 2 if #SMT_on else cpu@UOPS_EXECUTED.THREAD\\,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -1242,6 +1256,12 @@ "MetricGroup": "Fed;FetchBW", "MetricName": "tma_info_pipeline_fetch_mite" }, + { + "BriefDescription": "Average number of uops fetched from MS per cy= cle", + "MetricExpr": "IDQ.MS_UOPS / cpu@IDQ.MS_UOPS\\,cmask\\=3D1@", + "MetricGroup": "Fed;FetchLat;MicroSeq", + "MetricName": "tma_info_pipeline_fetch_ms" + }, { "BriefDescription": "Instructions per a microcode Assist invocatio= n", "MetricExpr": "INST_RETIRED.ANY / (FP_ASSIST.ANY + OTHER_ASSISTS.A= NY)", @@ -1258,7 +1278,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -1270,7 +1290,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -1279,7 +1299,7 @@ "MetricExpr": "64 * (UNC_M_CAS_COUNT.RD + UNC_M_CAS_COUNT.WR) / 1e= 9 / tma_info_system_time", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Giga Floating Point Operations Per Second", @@ -1475,12 +1495,13 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "min(2 * (MEM_INST_RETIRED.ALL_LOADS - MEM_LOAD_RETI= RED.FB_HIT - MEM_LOAD_RETIRED.L1_MISS) * 20 / 100, max(CYCLE_ACTIVITY.CYCLE= S_MEM_ANY - CYCLE_ACTIVITY.CYCLES_L1D_MISS, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%" }, { @@ -1517,7 +1538,7 @@ "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%" }, { @@ -1541,6 +1562,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s CPU dispatched uops on execution port for Load operations", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(UOPS_DISPATCHED_PORT.PORT_2 + UOPS_DISPATCHED_PORT= .PORT_3 + UOPS_DISPATCHED_PORT.PORT_7 - UOPS_DISPATCHED_PORT.PORT_4) / (2 *= tma_info_core_core_clks)", "MetricGroup": "TopdownL5;tma_L5_group;tma_ports_utilized_3m_group= ", "MetricName": "tma_load_op_utilization", @@ -1567,6 +1589,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 1 GB pages for= data load accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_load_stlb_miss * DTLB_LOAD_MISSES.WALK_COMPLETE= D_1G / (DTLB_LOAD_MISSES.WALK_COMPLETED_4K + DTLB_LOAD_MISSES.WALK_COMPLETE= D_2M_4M + DTLB_LOAD_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_load_stlb_mis= s_group", "MetricName": "tma_load_stlb_miss_1g", @@ -1575,6 +1598,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 2 or 4 MB page= s for data load accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_load_stlb_miss * DTLB_LOAD_MISSES.WALK_COMPLETE= D_2M_4M / (DTLB_LOAD_MISSES.WALK_COMPLETED_4K + DTLB_LOAD_MISSES.WALK_COMPL= ETED_2M_4M + DTLB_LOAD_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_load_stlb_mis= s_group", "MetricName": "tma_load_stlb_miss_2m", @@ -1583,6 +1607,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 4 KB pages for= data load accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_load_stlb_miss * DTLB_LOAD_MISSES.WALK_COMPLETE= D_4K / (DTLB_LOAD_MISSES.WALK_COMPLETED_4K + DTLB_LOAD_MISSES.WALK_COMPLETE= D_2M_4M + DTLB_LOAD_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_load_stlb_mis= s_group", "MetricName": "tma_load_stlb_miss_4k", @@ -1600,6 +1625,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU spent handling cache misses due to lock operations", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "(12 * max(0, MEM_INST_RETIRED.LOCK_LOADS - L2_RQSTS= .ALL_RFO) + MEM_INST_RETIRED.LOCK_LOADS / MEM_INST_RETIRED.ALL_STORES * (11= * L2_RQSTS.RFO_HIT + min(CPU_CLK_UNHALTED.THREAD, OFFCORE_REQUESTS_OUTSTAN= DING.CYCLES_WITH_DEMAND_RFO))) / tma_info_thread_clks", "MetricGroup": "LockCont;Offcore;TopdownL4;tma_L4_group;tma_issueR= FO;tma_l1_bound_group", "MetricName": "tma_lock_latency", @@ -1624,7 +1650,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%" }, { @@ -1633,7 +1659,7 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%" }, { @@ -1657,7 +1683,6 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU was retiring uops fetched by the Microcode Sequencer (MS) unit", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "UOPS_RETIRED.RETIRE_SLOTS / UOPS_ISSUED.ANY * IDQ.M= S_UOPS / tma_info_thread_slots", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_heavy_operatio= ns_group;tma_issueMC;tma_issueMS", "MetricName": "tma_microcode_sequencer", @@ -1667,6 +1692,7 @@ }, { "BriefDescription": "This metric represents fraction of cycles the= CPU was stalled due to Branch Resteers as a result of Branch Misprediction= at execution stage", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "BR_MISP_RETIRED.ALL_BRANCHES / (BR_MISP_RETIRED.ALL= _BRANCHES + MACHINE_CLEARS.COUNT) * INT_MISC.CLEAR_RESTEER_CYCLES / tma_inf= o_thread_clks", "MetricGroup": "BadSpec;BrMispredicts;BvMP;TopdownL4;tma_L4_group;= tma_branch_resteers_group;tma_issueBM", "MetricName": "tma_mispredicts_resteers", @@ -1721,6 +1747,7 @@ }, { "BriefDescription": "This metric represents the remaining light uo= ps fraction the CPU has executed - remaining means not covered by other sib= ling nodes", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(0, tma_light_operations - (tma_fp_arith + tma_m= emory_operations + tma_fused_instructions + tma_non_fused_branches))", "MetricGroup": "Pipeline;TopdownL3;tma_L3_group;tma_light_operatio= ns_group", "MetricName": "tma_other_light_ops", @@ -1730,6 +1757,7 @@ }, { "BriefDescription": "This metric estimates fraction of slots the C= PU was stalled due to other cases of misprediction (non-retired x86 branche= s or other types).", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(tma_branch_mispredicts * (1 - BR_MISP_RETIRED.A= LL_BRANCHES / (INT_MISC.CLEARS_COUNT - MACHINE_CLEARS.COUNT)), 0.0001)", "MetricGroup": "BrMispredicts;BvIO;TopdownL3;tma_L3_group;tma_bran= ch_mispredicts_group", "MetricName": "tma_other_mispredicts", @@ -1738,6 +1766,7 @@ }, { "BriefDescription": "This metric represents fraction of slots the = CPU has wasted due to Nukes (Machine Clears) not related to memory ordering= .", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "max(tma_machine_clears * (1 - MACHINE_CLEARS.MEMORY= _ORDERING / MACHINE_CLEARS.COUNT), 0.0001)", "MetricGroup": "BvIO;Machine_Clears;TopdownL3;tma_L3_group;tma_mac= hine_clears_group", "MetricName": "tma_other_nukes", @@ -1818,6 +1847,7 @@ }, { "BriefDescription": "This metric estimates fraction of cycles the = CPU performance was potentially limited due to Core computation issues (non= divider-related)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "((tma_ports_utilized_0 * tma_info_thread_clks + (EX= E_ACTIVITY.1_PORTS_UTIL + tma_retiring * EXE_ACTIVITY.2_PORTS_UTIL)) / tma_= info_thread_clks if ARITH.DIVIDER_ACTIVE < CYCLE_ACTIVITY.STALLS_TOTAL - CY= CLE_ACTIVITY.STALLS_MEM_ANY else (EXE_ACTIVITY.1_PORTS_UTIL + tma_retiring = * EXE_ACTIVITY.2_PORTS_UTIL) / tma_info_thread_clks)", "MetricGroup": "PortsUtil;TopdownL3;tma_L3_group;tma_core_bound_gr= oup", "MetricName": "tma_ports_utilization", @@ -1923,7 +1953,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%" }, { @@ -1980,6 +2010,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 1 GB pages for= data store accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_store_stlb_miss * DTLB_STORE_MISSES.WALK_COMPLE= TED_1G / (DTLB_STORE_MISSES.WALK_COMPLETED_4K + DTLB_STORE_MISSES.WALK_COMP= LETED_2M_4M + DTLB_STORE_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_store_stlb_mi= ss_group", "MetricName": "tma_store_stlb_miss_1g", @@ -1988,6 +2019,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 2 or 4 MB page= s for data store accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_store_stlb_miss * DTLB_STORE_MISSES.WALK_COMPLE= TED_2M_4M / (DTLB_STORE_MISSES.WALK_COMPLETED_4K + DTLB_STORE_MISSES.WALK_C= OMPLETED_2M_4M + DTLB_STORE_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_store_stlb_mi= ss_group", "MetricName": "tma_store_stlb_miss_2m", @@ -1996,6 +2028,7 @@ }, { "BriefDescription": "This metric estimates the fraction of cycles = to walk the memory paging structures to cache translation of 4 KB pages for= data store accesses.", + "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "tma_store_stlb_miss * DTLB_STORE_MISSES.WALK_COMPLE= TED_4K / (DTLB_STORE_MISSES.WALK_COMPLETED_4K + DTLB_STORE_MISSES.WALK_COMP= LETED_2M_4M + DTLB_STORE_MISSES.WALK_COMPLETED_1G)", "MetricGroup": "MemoryTLB;TopdownL6;tma_L6_group;tma_store_stlb_mi= ss_group", "MetricName": "tma_store_stlb_miss_4k", --=20 2.51.0.rc1.167.g924127e9c0-goog From nobody Sat Oct 4 08:08:45 2025 Received: from mail-pl1-f202.google.com (mail-pl1-f202.google.com [209.85.214.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 858513451AA for ; Mon, 18 Aug 2025 19:06:31 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.202 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543995; cv=none; b=eq51dcs4hmtwFyXCj63sB4f2rJ79VCoPn2zCxx1/KT8K/sj1cf/I9IBRAp1QxGECgeh01J3DHAxZyv5v3gloeIZ7IDBHF6EyZ7Ny/CGkoV7MFlQu24dq8a/HK3uVrXpciLjkxu65gdkIyVfsRTfe46ZNtAJIeXZBhIkzDUiQp/o= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755543995; c=relaxed/simple; bh=UsXw9x29ZbNtGM96teVkQJv0iivIpANKZ4rBeHTFkLc=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Content-Type; b=mF+EP02LHbWqbQF0fuv3wCLNAW3MC1n0/MYMnZ+lXlkAc/ha8NOegMbmJWe1yAmJAE/k6VtEKZ4yNs+pP8FkXzT+qAvptVAzpCuslnJ32isPBHKJBG5EFy2uzY3tYsl3jxf59eT8WTBcFlkbwlZR/H/L2U9DuWdyiWpqalmGw58= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=TXGtJY4s; arc=none smtp.client-ip=209.85.214.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--irogers.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="TXGtJY4s" Received: by mail-pl1-f202.google.com with SMTP id d9443c01a7336-2445823bc21so112559265ad.3 for ; Mon, 18 Aug 2025 12:06:31 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1755543991; x=1756148791; darn=vger.kernel.org; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:from:to:cc:subject:date:message-id :reply-to; bh=OadSD3Pb1UTaC9Y8QE9413VXg/3egXRdL1KX4X+y1NQ=; b=TXGtJY4s0oCLwNsC+O2ZFd7wPmSbaXaU+Vm1KIjIt8c2nxJEE0wX6jyQSxRDLqe4AR r5lHcZuEXfKuuQKDaLk+t//V+1nRAsFaV3Zshld8GgwDUTn76/t3lR60nJbLDzFFai// jDdZHD/wicAyAm1WW/lrbyxsBbZ5npxPUyAFWYBFo++d1TKM78OqIJ0/9f4ex0lHuc61 Ml8sbf77nziQ1E3+r2HvmlRactqWx8pjq/RcNzNu4rHUzV/P5tgcBC62ozekU93/XFD/ B9HuuGiI07rDoqrRSnEs2Q8ZrQO6kyTQ1GBOx23Obnzkljuhu9xGC7OyI4Fjp8ixjWFf T6Tw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1755543991; x=1756148791; h=content-transfer-encoding:to:from:subject:message-id:references :mime-version:in-reply-to:date:x-gm-message-state:from:to:cc:subject :date:message-id:reply-to; bh=OadSD3Pb1UTaC9Y8QE9413VXg/3egXRdL1KX4X+y1NQ=; b=rOkmD/k2VfsUFQP7B/jCV1QI3fNejoc+FPcB7hv37y6qXU7lkrhMCdkgECm3FZgi9A 7gJhiUu0LBv9DZoESGFU/NgDsd5i7OgCpLRew37QvZ8G9u1mgtevq7rnJf2ZP8EX0M53 9QOP/hJJIJJAvUr9uGFH/vvWQyZXiBiuHRS1N7uVu864Ahx8gKkx+R2QYXNXGBJCsBvV tL2lHi7oJqZRMbSykgQLiRC61FXRKOTnzDCLxpZzerAzC6PJDqf17kb9IN5cBGKRwXo9 Klc0UglHrUa9HYITk9Fu3CziMVPq+NYjW1h9jzLMrMfrfYn0rLLYBuN+mq/pQcB4gVDe /LNw== X-Forwarded-Encrypted: i=1; AJvYcCVXKlXmycm6r/F3LthVkDYXLEVQbiSmMJLI7ZvftB/61VurjLv3JWcSo26pggTGONZDu2dc+K3M2pjSY7w=@vger.kernel.org X-Gm-Message-State: AOJu0Yzg8Yj4zo0cjlWlDDGa3MUp+jSUwL5tUabXuu4WaRMuZYN2og4L Cq5lcTSI4MIGwVJXE8Uy1SJnY/Jp5Hn1oSg2aCQ71vfjNQy6lyEiaVvflzHLW4Dng87cFU/DwAg UgrkLZFHAOg== X-Google-Smtp-Source: AGHT+IFg33d5dO+ENVrLNt1kWtvSPQy3frFazMkPPlbi8VHrevR4RYgEp3gx9aTo45PTs5mFZwpAiEKJ/JxL X-Received: from plblh13.prod.google.com ([2002:a17:903:290d:b0:240:52d7:e88a]) (user=irogers job=prod-delivery.src-stubby-dispatcher) by 2002:a17:903:1249:b0:235:ec11:f0ee with SMTP id d9443c01a7336-2449cf92875mr5242365ad.14.1755543990751; Mon, 18 Aug 2025 12:06:30 -0700 (PDT) Date: Mon, 18 Aug 2025 12:04:16 -0700 In-Reply-To: <20250818190416.145274-1-irogers@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250818190416.145274-1-irogers@google.com> X-Mailer: git-send-email 2.51.0.rc1.167.g924127e9c0-goog Message-ID: <20250818190416.145274-21-irogers@google.com> Subject: [PATCH v2 20/20] perf vendor events: Update tigerlake metrics From: Ian Rogers To: Peter Zijlstra , Ingo Molnar , Arnaldo Carvalho de Melo , Namhyung Kim , Mark Rutland , Alexander Shishkin , Jiri Olsa , Ian Rogers , Adrian Hunter , Kan Liang , "=?UTF-8?q?Andreas=20F=C3=A4rber?=" , Manivannan Sadhasivam , Caleb Biggers , Thomas Falcon , Weilin Wang , linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org, linux-actions@lists.infradead.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Update metrics from TMA 5.0 to 5.1. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon --- .../arch/x86/tigerlake/tgl-metrics.json | 97 +++++++++++-------- 1 file changed, 56 insertions(+), 41 deletions(-) diff --git a/tools/perf/pmu-events/arch/x86/tigerlake/tgl-metrics.json b/to= ols/perf/pmu-events/arch/x86/tigerlake/tgl-metrics.json index 2db7a70f7a07..908da985c594 100644 --- a/tools/perf/pmu-events/arch/x86/tigerlake/tgl-metrics.json +++ b/tools/perf/pmu-events/arch/x86/tigerlake/tgl-metrics.json @@ -1,63 +1,63 @@ [ { "BriefDescription": "C10 residency percent per package", - "MetricExpr": "cstate_pkg@c10\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c10\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C10_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C2 residency percent per package", - "MetricExpr": "cstate_pkg@c2\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c2\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C2_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C3 residency percent per package", - "MetricExpr": "cstate_pkg@c3\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c3\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C3_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per core", - "MetricExpr": "cstate_core@c6\\-residency@ / TSC", + "MetricExpr": "cstate_core@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C6 residency percent per package", - "MetricExpr": "cstate_pkg@c6\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c6\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C6_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per core", - "MetricExpr": "cstate_core@c7\\-residency@ / TSC", + "MetricExpr": "cstate_core@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Core_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C7 residency percent per package", - "MetricExpr": "cstate_pkg@c7\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c7\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C7_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C8 residency percent per package", - "MetricExpr": "cstate_pkg@c8\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c8\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C8_Pkg_Residency", "ScaleUnit": "100%" }, { "BriefDescription": "C9 residency percent per package", - "MetricExpr": "cstate_pkg@c9\\-residency@ / TSC", + "MetricExpr": "cstate_pkg@c9\\-residency@ / msr@tsc@", "MetricGroup": "Power", "MetricName": "C9_Pkg_Residency", "ScaleUnit": "100%" @@ -85,7 +85,6 @@ }, { "BriefDescription": "This metric estimates how often memory load a= ccesses were aliased by preceding stores (in program order) with a 4K addre= ss offset", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "LD_BLOCKS_PARTIAL.ADDRESS_ALIAS / tma_info_thread_c= lks", "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", "MetricName": "tma_4k_aliasing", @@ -134,6 +133,7 @@ }, { "BriefDescription": "Total pipeline cost of instruction fetch rela= ted bottlenecks by large code footprint programs (i-side cache; TLB and BTB= misses)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * tma_fetch_latency * (tma_itlb_misses + tma_ic= ache_misses + tma_unknown_branches) / (tma_branch_resteers + tma_dsb_switch= es + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches)", "MetricGroup": "BigFootprint;BvBC;Fed;Frontend;IcMiss;MemoryTLB", "MetricName": "tma_bottleneck_big_code", @@ -147,40 +147,45 @@ "MetricThreshold": "tma_bottleneck_branching_overhead > 5", "PublicDescription": "Total pipeline cost of instructions used for= program control-flow - a subset of the Retiring category in TMA. Examples = include function calls; loops and alignments. (A lower bound)" }, + { + "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", + "MetricGroup": "BvCB;Cor;tma_issueComp", + "MetricName": "tma_bottleneck_compute_bound_est", + "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", + "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " + }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Bandwidth related bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_bandwidth / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_b= ound * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_= l3_bound + tma_store_bound)) * (tma_sq_full / (tma_contested_accesses + tma= _data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound * (tm= a_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound += tma_store_bound)) * (tma_fb_full / (tma_4k_aliasing + tma_dtlb_load + tma_= fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_loads + = tma_store_fwd_blk)))", "MetricGroup": "BvMB;Mem;MemoryBW;Offcore;tma_issueBW", - "MetricName": "tma_bottleneck_cache_memory_bandwidth", - "MetricThreshold": "tma_bottleneck_cache_memory_bandwidth > 20", + "MetricName": "tma_bottleneck_data_cache_memory_bandwidth", + "MetricThreshold": "tma_bottleneck_data_cache_memory_bandwidth > 2= 0", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Bandwidth related bottlenecks. Related metrics: tma_fb_full, tma_info_= system_dram_bw_use, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Total pipeline cost of external Memory- or Ca= che-Latency related bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_dram_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) *= (tma_mem_latency / (tma_mem_bandwidth + tma_mem_latency)) + tma_memory_bou= nd * (tma_l3_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3= _bound + tma_store_bound)) * (tma_l3_hit_latency / (tma_contested_accesses = + tma_data_sharing + tma_l3_hit_latency + tma_sq_full)) + tma_memory_bound = * tma_l2_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bou= nd + tma_store_bound) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound = + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_l1_= latency_dependency / (tma_4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l= 1_latency_dependency + tma_lock_latency + tma_split_loads + tma_store_fwd_b= lk)) + tma_memory_bound * (tma_l1_bound / (tma_dram_bound + tma_l1_bound + = tma_l2_bound + tma_l3_bound + tma_store_bound)) * (tma_lock_latency / (tma_= 4k_aliasing + tma_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma= _lock_latency + tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * = (tma_l1_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_boun= d + tma_store_bound)) * (tma_split_loads / (tma_4k_aliasing + tma_dtlb_load= + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + tma_split_l= oads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound / (tma_dra= m_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound)) * = (tma_split_stores / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency + tma_streaming_stores)) + tma_memory_bound * (tma_stor= e_bound / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tm= a_store_bound)) * (tma_store_latency / (tma_dtlb_store + tma_false_sharing = + tma_split_stores + tma_store_latency + tma_streaming_stores)))", "MetricGroup": "BvML;Mem;MemoryLat;Offcore;tma_issueLat", - "MetricName": "tma_bottleneck_cache_memory_latency", - "MetricThreshold": "tma_bottleneck_cache_memory_latency > 20", + "MetricName": "tma_bottleneck_data_cache_memory_latency", + "MetricThreshold": "tma_bottleneck_data_cache_memory_latency > 20", "PublicDescription": "Total pipeline cost of external Memory- or C= ache-Latency related bottlenecks. Related metrics: tma_l3_hit_latency, tma_= mem_latency" }, - { - "BriefDescription": "Total pipeline cost when the execution is com= pute-bound - an estimation", - "MetricExpr": "100 * (tma_core_bound * tma_divider / (tma_divider = + tma_ports_utilization + tma_serializing_operation) + tma_core_bound * (tm= a_ports_utilization / (tma_divider + tma_ports_utilization + tma_serializin= g_operation)) * (tma_ports_utilized_3m / (tma_ports_utilized_0 + tma_ports_= utilized_1 + tma_ports_utilized_2 + tma_ports_utilized_3m)))", - "MetricGroup": "BvCB;Cor;tma_issueComp", - "MetricName": "tma_bottleneck_compute_bound_est", - "MetricThreshold": "tma_bottleneck_compute_bound_est > 20", - "PublicDescription": "Total pipeline cost when the execution is co= mpute-bound - an estimation. Covers Core Bound when High ILP as well as whe= n long-latency execution units are busy. Related metrics: " - }, { "BriefDescription": "Total pipeline cost of instruction fetch band= width related bottlenecks (when the front-end could not sustain operations = delivery to the back-end)", - "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + 10 * tma_microcode_seq= uencer * tma_other_mispredicts / tma_branch_mispredicts * tma_mispredicts_r= esteers) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_br= anches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tm= a_itlb_misses + tma_lcp + tma_ms_switches) + tma_fetch_bandwidth * tma_ms /= (tma_dsb + tma_lsd + tma_mite + tma_ms))) - tma_bottleneck_big_code", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_frontend_bound - (1 - 10 * tma_microcode= _sequencer * tma_other_mispredicts / tma_branch_mispredicts) * tma_fetch_la= tency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switches = + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches) - tma_mi= crocode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer) *= (tma_assists / tma_microcode_sequencer) * (tma_fetch_latency * (tma_ms_swi= tches + tma_branch_resteers * (tma_clears_resteers + 10 * tma_microcode_seq= uencer * tma_other_mispredicts / tma_branch_mispredicts * tma_mispredicts_r= esteers) / (tma_clears_resteers + tma_mispredicts_resteers + tma_unknown_br= anches)) / (tma_branch_resteers + tma_dsb_switches + tma_icache_misses + tm= a_itlb_misses + tma_lcp + tma_ms_switches) + tma_ms)) - tma_bottleneck_big_= code", "MetricGroup": "BvFB;Fed;FetchBW;Frontend", "MetricName": "tma_bottleneck_instruction_fetch_bw", "MetricThreshold": "tma_bottleneck_instruction_fetch_bw > 20" }, { "BriefDescription": "Total pipeline cost of irregular execution (e= .g", - "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cle= ars_resteers + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_b= ranch_mispredicts * tma_mispredicts_resteers) / (tma_clears_resteers + tma_= mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_= dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switc= hes) + tma_fetch_bandwidth * tma_ms / (tma_dsb + tma_lsd + tma_mite + tma_m= s)) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_branch_mis= predicts * tma_branch_mispredicts + tma_machine_clears * tma_other_nukes / = tma_other_nukes + tma_core_bound * (tma_serializing_operation + tma_core_bo= und * RS_EVENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_utilized_0)= / (tma_divider + tma_ports_utilization + tma_serializing_operation) + tma_= microcode_sequencer / (tma_few_uops_instructions + tma_microcode_sequencer)= * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 * (tma_microcode_sequencer / (tma_few_uops_inst= ructions + tma_microcode_sequencer) * (tma_assists / tma_microcode_sequence= r) * (tma_fetch_latency * (tma_ms_switches + tma_branch_resteers * (tma_cle= ars_resteers + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma_b= ranch_mispredicts * tma_mispredicts_resteers) / (tma_clears_resteers + tma_= mispredicts_resteers + tma_unknown_branches)) / (tma_branch_resteers + tma_= dsb_switches + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switc= hes) + tma_ms) + 10 * tma_microcode_sequencer * tma_other_mispredicts / tma= _branch_mispredicts * tma_branch_mispredicts + tma_machine_clears * tma_oth= er_nukes / tma_other_nukes + tma_core_bound * (tma_serializing_operation + = tma_core_bound * RS_EVENTS.EMPTY_CYCLES / tma_info_thread_clks * tma_ports_= utilized_0) / (tma_divider + tma_ports_utilization + tma_serializing_operat= ion) + tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode= _sequencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operation= s)", "MetricGroup": "Bad;BvIO;Cor;Ret;tma_issueMS", "MetricName": "tma_bottleneck_irregular_overhead", "MetricThreshold": "tma_bottleneck_irregular_overhead > 10", @@ -188,6 +193,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Address Transla= tion related bottlenecks (data-side TLBs)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_l1_bound / max(tma_m= emory_bound, tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + = tma_store_bound)) * (tma_dtlb_load / max(tma_l1_bound, tma_4k_aliasing + tm= a_dtlb_load + tma_fb_full + tma_l1_latency_dependency + tma_lock_latency + = tma_split_loads + tma_store_fwd_blk)) + tma_memory_bound * (tma_store_bound= / (tma_dram_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store= _bound)) * (tma_dtlb_store / (tma_dtlb_store + tma_false_sharing + tma_spli= t_stores + tma_store_latency + tma_streaming_stores)))", "MetricGroup": "BvMT;Mem;MemoryTLB;Offcore;tma_issueTLB", "MetricName": "tma_bottleneck_memory_data_tlbs", @@ -196,6 +202,7 @@ }, { "BriefDescription": "Total pipeline cost of Memory Synchronization= related bottlenecks (data transfers and coherency updates across processor= s)", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_memory_bound * (tma_l3_bound / (tma_dram= _bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * (t= ma_contested_accesses + tma_data_sharing) / (tma_contested_accesses + tma_d= ata_sharing + tma_l3_hit_latency + tma_sq_full) + tma_store_bound / (tma_dr= am_bound + tma_l1_bound + tma_l2_bound + tma_l3_bound + tma_store_bound) * = tma_false_sharing / (tma_dtlb_store + tma_false_sharing + tma_split_stores = + tma_store_latency + tma_streaming_stores - tma_store_latency)) + tma_mach= ine_clears * (1 - tma_other_nukes / tma_other_nukes))", "MetricGroup": "BvMS;LockCont;Mem;Offcore;tma_issueSyncxn", "MetricName": "tma_bottleneck_memory_synchronization", @@ -204,6 +211,7 @@ }, { "BriefDescription": "Total pipeline cost of Branch Misprediction r= elated bottlenecks", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (1 - 10 * tma_microcode_sequencer * tma_other= _mispredicts / tma_branch_mispredicts) * (tma_branch_mispredicts + tma_fetc= h_latency * tma_mispredicts_resteers / (tma_branch_resteers + tma_dsb_switc= hes + tma_icache_misses + tma_itlb_misses + tma_lcp + tma_ms_switches))", "MetricGroup": "Bad;BadSpec;BrMispredicts;BvMP;tma_issueBM", "MetricName": "tma_bottleneck_mispredictions", @@ -212,7 +220,8 @@ }, { "BriefDescription": "Total pipeline cost of remaining bottlenecks = in the back-end", - "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_cache_me= mory_bandwidth + tma_bottleneck_cache_memory_latency + tma_bottleneck_memor= y_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottleneck_comput= e_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_branching_= overhead + tma_bottleneck_useful_work)", + "MetricConstraint": "NO_GROUP_EVENTS", + "MetricExpr": "100 - (tma_bottleneck_big_code + tma_bottleneck_ins= truction_fetch_bw + tma_bottleneck_mispredictions + tma_bottleneck_data_cac= he_memory_bandwidth + tma_bottleneck_data_cache_memory_latency + tma_bottle= neck_memory_data_tlbs + tma_bottleneck_memory_synchronization + tma_bottlen= eck_compute_bound_est + tma_bottleneck_irregular_overhead + tma_bottleneck_= branching_overhead + tma_bottleneck_useful_work)", "MetricGroup": "BvOB;Cor;Offcore", "MetricName": "tma_bottleneck_other_bottlenecks", "MetricThreshold": "tma_bottleneck_other_bottlenecks > 20", @@ -220,6 +229,7 @@ }, { "BriefDescription": "Total pipeline cost of \"useful operations\" = - the portion of Retiring category not covered by Branching_Overhead nor Ir= regular_Overhead.", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_retiring - (BR_INST_RETIRED.ALL_BRANCHES= + 2 * BR_INST_RETIRED.NEAR_CALL + INST_RETIRED.NOP) / tma_info_thread_slot= s - tma_microcode_sequencer / (tma_few_uops_instructions + tma_microcode_se= quencer) * (tma_assists / tma_microcode_sequencer) * tma_heavy_operations)", "MetricGroup": "BvUW;Ret", "MetricName": "tma_bottleneck_useful_work", @@ -427,7 +437,7 @@ "MetricGroup": "BvMB;MemoryBW;TopdownL4;tma_L4_group;tma_issueBW;t= ma_issueSL;tma_issueSmSt;tma_l1_bound_group", "MetricName": "tma_fb_full", "MetricThreshold": "tma_fb_full > 0.3", - "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_cache_memory_bandwidth, = tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store_late= ncy, tma_streaming_stores", + "PublicDescription": "This metric does a *rough estimation* of how= often L1D Fill Buffer unavailability limited additional L1D miss memory ac= cess requests to proceed. The higher the metric value; the deeper the memor= y hierarchy level the misses are satisfied from (metric values >1 are valid= ). Often it hints on approaching bandwidth limits (to L2 cache; L3 cache or= external memory). Related metrics: tma_bottleneck_data_cache_memory_bandwi= dth, tma_info_system_dram_bw_use, tma_mem_bandwidth, tma_sq_full, tma_store= _latency, tma_streaming_stores", "ScaleUnit": "100%" }, { @@ -619,6 +629,7 @@ }, { "BriefDescription": "Total pipeline cost of DSB (uop cache) hits -= subset of the Instruction_Fetch_BW Bottleneck", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "100 * (tma_frontend_bound * (tma_fetch_bandwidth / = (tma_fetch_bandwidth + tma_fetch_latency)) * (tma_dsb / (tma_dsb + tma_lsd = + tma_mite + tma_ms)))", "MetricGroup": "DSB;Fed;FetchBW;tma_issueFB", "MetricName": "tma_info_botlnk_l2_dsb_bandwidth", @@ -1074,7 +1085,7 @@ "MetricName": "tma_info_memory_tlb_store_stlb_mpki" }, { - "BriefDescription": "", + "BriefDescription": "Mem;Backend;CacheHits", "MetricExpr": "UOPS_EXECUTED.THREAD / (UOPS_EXECUTED.CORE_CYCLES_G= E_1 / 2 if #SMT_on else cpu@UOPS_EXECUTED.THREAD\\,cmask\\=3D1@)", "MetricGroup": "Cor;Pipeline;PortsUtil;SMT", "MetricName": "tma_info_pipeline_execute" @@ -1097,6 +1108,12 @@ "MetricGroup": "Fed;FetchBW", "MetricName": "tma_info_pipeline_fetch_mite" }, + { + "BriefDescription": "Average number of uops fetched from MS per cy= cle", + "MetricExpr": "IDQ.MS_UOPS / cpu@IDQ.MS_UOPS\\,cmask\\=3D1@", + "MetricGroup": "Fed;FetchLat;MicroSeq", + "MetricName": "tma_info_pipeline_fetch_ms" + }, { "BriefDescription": "Instructions per a microcode Assist invocatio= n", "MetricExpr": "INST_RETIRED.ANY / ASSISTS.ANY", @@ -1113,7 +1130,7 @@ }, { "BriefDescription": "Measured Average Core Frequency for unhalted = processors [GHz]", - "MetricExpr": "tma_info_system_turbo_utilization * TSC / 1e9 / tma= _info_system_time", + "MetricExpr": "tma_info_system_turbo_utilization * msr@tsc@ / 1e9 = / tma_info_system_time", "MetricGroup": "Power;Summary", "MetricName": "tma_info_system_core_frequency" }, @@ -1125,7 +1142,7 @@ }, { "BriefDescription": "Average number of utilized CPUs", - "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / TSC", + "MetricExpr": "CPU_CLK_UNHALTED.REF_TSC / msr@tsc@", "MetricGroup": "Summary", "MetricName": "tma_info_system_cpus_utilized" }, @@ -1134,7 +1151,7 @@ "MetricExpr": "64 * (UNC_ARB_TRK_REQUESTS.ALL + UNC_ARB_COH_TRK_RE= QUESTS.ALL) / 1e6 / tma_info_system_time / 1e3", "MetricGroup": "HPC;MemOffcore;MemoryBW;SoC;tma_issueBW", "MetricName": "tma_info_system_dram_bw_use", - "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_cache_memory_ban= dwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" + "PublicDescription": "Average external Memory Bandwidth Use for re= ads and writes [GB / sec]. Related metrics: tma_bottleneck_data_cache_memor= y_bandwidth, tma_fb_full, tma_mem_bandwidth, tma_sq_full" }, { "BriefDescription": "Giga Floating Point Operations Per Second", @@ -1165,6 +1182,7 @@ }, { "BriefDescription": "Average number of parallel data read requests= to external memory", + "MetricConstraint": "NO_GROUP_EVENTS", "MetricExpr": "UNC_ARB_DAT_OCCUPANCY.RD / UNC_ARB_DAT_OCCUPANCY.RD= @cmask\\=3D1@", "MetricGroup": "Mem;MemoryBW;SoC", "MetricName": "tma_info_system_mem_parallel_reads", @@ -1316,12 +1334,12 @@ "ScaleUnit": "100%" }, { - "BriefDescription": "This metric([SKL+] roughly; [LNL]) estimates = fraction of cycles with demand load accesses that hit the L1D cache", + "BriefDescription": "This metric ([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache", "MetricExpr": "min(2 * (MEM_INST_RETIRED.ALL_LOADS - MEM_LOAD_RETI= RED.FB_HIT - MEM_LOAD_RETIRED.L1_MISS) * 20 / 100, max(CYCLE_ACTIVITY.CYCLE= S_MEM_ANY - CYCLE_ACTIVITY.CYCLES_L1D_MISS, 0)) / tma_info_thread_clks", "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_l1_bound= _group", "MetricName": "tma_l1_latency_dependency", "MetricThreshold": "tma_l1_latency_dependency > 0.1 & (tma_l1_boun= d > 0.1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric([SKL+] roughly; [LNL]) estimates= fraction of cycles with demand load accesses that hit the L1D cache. The s= hort latency of the L1D cache may be exposed in pointer-chasing memory acce= ss patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", + "PublicDescription": "This metric ([SKL+] roughly; [LNL]) estimate= s fraction of cycles with demand load accesses that hit the L1D cache. The = short latency of the L1D cache may be exposed in pointer-chasing memory acc= ess patterns as an example. Sample with: MEM_LOAD_RETIRED.L1_HIT", "ScaleUnit": "100%" }, { @@ -1345,7 +1363,6 @@ }, { "BriefDescription": "This metric estimates how often the CPU was s= talled due to loads accesses to L3 cache or contended with a sibling Core", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "(CYCLE_ACTIVITY.STALLS_L2_MISS - CYCLE_ACTIVITY.STA= LLS_L3_MISS) / tma_info_thread_clks", "MetricGroup": "CacheHits;MemoryBound;TmaL3mem;TopdownL3;tma_L3_gr= oup;tma_memory_bound_group", "MetricName": "tma_l3_bound", @@ -1359,7 +1376,7 @@ "MetricGroup": "BvML;MemoryLat;TopdownL4;tma_L4_group;tma_issueLat= ;tma_l3_bound_group", "MetricName": "tma_l3_hit_latency", "MetricThreshold": "tma_l3_hit_latency > 0.1 & (tma_l3_bound > 0.0= 5 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_cache_memory_latency, tma_mem_latency", + "PublicDescription": "This metric estimates fraction of cycles wit= h demand load accesses that hit the L3 cache under unloaded scenarios (poss= ibly L3 latency limited). Avoiding private cache misses (i.e. L2 misses/L3= hits) will improve the latency; reduce contention with sibling physical co= res and increase performance. Note the value of this node may overlap with= its siblings. Sample with: MEM_LOAD_RETIRED.L3_HIT_PS. Related metrics: tm= a_bottleneck_data_cache_memory_latency, tma_mem_latency", "ScaleUnit": "100%" }, { @@ -1465,7 +1482,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_d= ram_bound_group;tma_issueBW", "MetricName": "tma_mem_bandwidth", "MetricThreshold": "tma_mem_bandwidth > 0.2 & (tma_dram_bound > 0.= 1 & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_cache_memory_bandwidth,= tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", + "PublicDescription": "This metric estimates fraction of cycles whe= re the core's performance was likely hurt due to approaching bandwidth limi= ts of external memory - DRAM ([SPR-HBM] and/or HBM). The underlying heuris= tic assumes that a similar off-core traffic is generated by all IA cores. T= his metric does not aggregate non-data-read requests by this logical proces= sor; requests from other IA Logical Processors/Physical Cores/sockets; or o= ther non-IA devices like GPU; hence the maximum external memory bandwidth l= imits may or may not be approached when this metric is flagged (see Uncore = counters for that). Related metrics: tma_bottleneck_data_cache_memory_bandw= idth, tma_fb_full, tma_info_system_dram_bw_use, tma_sq_full", "ScaleUnit": "100%" }, { @@ -1474,7 +1491,7 @@ "MetricGroup": "BvML;MemoryLat;Offcore;TopdownL4;tma_L4_group;tma_= dram_bound_group;tma_issueLat", "MetricName": "tma_mem_latency", "MetricThreshold": "tma_mem_latency > 0.1 & (tma_dram_bound > 0.1 = & (tma_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_cache_memory_latency, tma_l3_hit_latency= ", + "PublicDescription": "This metric estimates fraction of cycles whe= re the performance was likely hurt due to latency from external memory - DR= AM ([SPR-HBM] and/or HBM). This metric does not aggregate requests from ot= her Logical Processors/Physical Cores/sockets (see Uncore counters for that= ). Related metrics: tma_bottleneck_data_cache_memory_latency, tma_l3_hit_la= tency", "ScaleUnit": "100%" }, { @@ -1542,7 +1559,7 @@ }, { "BriefDescription": "This metric represents Core fraction of cycle= s in which CPU was likely limited due to the Microcode Sequencer (MS) unit = - see Microcode_Sequencer node for details.", - "MetricExpr": "cpu@IDQ.MS_UOPS\\,cmask\\=3D1@ / tma_info_core_core= _clks / 2", + "MetricExpr": "cpu@IDQ.MS_UOPS\\,cmask\\=3D1@ / tma_info_core_core= _clks / 3.3", "MetricGroup": "MicroSeq;TopdownL3;tma_L3_group;tma_fetch_bandwidt= h_group", "MetricName": "tma_ms", "MetricThreshold": "tma_ms > 0.05 & tma_fetch_bandwidth > 0.2", @@ -1676,7 +1693,7 @@ { "BriefDescription": "This category represents fraction of slots ut= ilized by useful work i.e. issued uops that eventually get retired", "DefaultMetricgroupName": "TopdownL1", - "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound) + 0 * tma_info_= thread_slots", + "MetricExpr": "topdown\\-retiring / (topdown\\-fe\\-bound + topdow= n\\-bad\\-spec + topdown\\-retiring + topdown\\-be\\-bound)", "MetricGroup": "BvUW;Default;TmaL1;TopdownL1;tma_L1_group", "MetricName": "tma_retiring", "MetricThreshold": "tma_retiring > 0.7 | tma_heavy_operations > 0.= 1", @@ -1713,7 +1730,6 @@ }, { "BriefDescription": "This metric represents rate of split store ac= cesses", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "MEM_INST_RETIRED.SPLIT_STORES / tma_info_core_core_= clks", "MetricGroup": "TopdownL4;tma_L4_group;tma_issueSpSt;tma_store_bou= nd_group", "MetricName": "tma_split_stores", @@ -1727,7 +1743,7 @@ "MetricGroup": "BvMB;MemoryBW;Offcore;TopdownL4;tma_L4_group;tma_i= ssueBW;tma_l3_bound_group", "MetricName": "tma_sq_full", "MetricThreshold": "tma_sq_full > 0.3 & (tma_l3_bound > 0.05 & (tm= a_memory_bound > 0.2 & tma_backend_bound > 0.2))", - "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, tma_m= em_bandwidth", + "PublicDescription": "This metric measures fraction of cycles wher= e the Super Queue (SQ) was full taking into account all request-types and b= oth hardware SMT threads (Logical Processors). Related metrics: tma_bottlen= eck_data_cache_memory_bandwidth, tma_fb_full, tma_info_system_dram_bw_use, = tma_mem_bandwidth", "ScaleUnit": "100%" }, { @@ -1741,7 +1757,6 @@ }, { "BriefDescription": "This metric roughly estimates fraction of cyc= les when the memory subsystem had loads blocked since they could not forwar= d data from earlier (in program order) overlapping stores", - "MetricConstraint": "NO_GROUP_EVENTS_NMI", "MetricExpr": "13 * LD_BLOCKS.STORE_FORWARD / tma_info_thread_clks= ", "MetricGroup": "TopdownL4;tma_L4_group;tma_l1_bound_group", "MetricName": "tma_store_fwd_blk", --=20 2.51.0.rc1.167.g924127e9c0-goog