frontend_bound metrics was miscalculated due to different scaling in
a couple of metrics it depends on. Change the scaling to match with
AmpereOne.

Fixes: 16438b652b46 ("perf vendor events arm64 AmpereOneX: Add core PMU events and metrics")
Signed-off-by: Ilkka Koskinen <ilkka@os.amperecomputing.com>
---
 .../arch/arm64/ampere/ampereonex/metrics.json          | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/tools/perf/pmu-events/arch/arm64/ampere/ampereonex/metrics.json b/tools/perf/pmu-events/arch/arm64/ampere/ampereonex/metrics.json
index c5d1d22bd034..5228f94a793f 100644
--- a/tools/perf/pmu-events/arch/arm64/ampere/ampereonex/metrics.json
+++ b/tools/perf/pmu-events/arch/arm64/ampere/ampereonex/metrics.json
@@ -229,19 +229,19 @@
     },
     {
         "MetricName": "slots_lost_misspeculation_fraction",
-        "MetricExpr": "(OP_SPEC - OP_RETIRED) / (CPU_CYCLES * #slots)",
+        "MetricExpr": "100 * (OP_SPEC - OP_RETIRED) / (CPU_CYCLES * #slots)",
         "BriefDescription": "Fraction of slots lost due to misspeculation",
         "DefaultMetricgroupName": "TopdownL1",
         "MetricGroup": "Default;TopdownL1",
-        "ScaleUnit": "100percent of slots"
+        "ScaleUnit": "1percent of slots"
     },
     {
         "MetricName": "retired_fraction",
-        "MetricExpr": "OP_RETIRED / (CPU_CYCLES * #slots)",
+        "MetricExpr": "100 * OP_RETIRED / (CPU_CYCLES * #slots)",
         "BriefDescription": "Fraction of slots retiring, useful work",
         "DefaultMetricgroupName": "TopdownL1",
         "MetricGroup": "Default;TopdownL1",
-        "ScaleUnit": "100percent of slots"
+        "ScaleUnit": "1percent of slots"
     },
     {
         "MetricName": "backend_core",
@@ -266,7 +266,7 @@
     },
     {
         "MetricName": "frontend_bandwidth",
-        "MetricExpr": "frontend_bound - frontend_latency",
+        "MetricExpr": "frontend_bound - 100 * frontend_latency",
         "BriefDescription": "Fraction of slots the CPU did not dispatch at full bandwidth - able to dispatch partial slots only (1, 2, or 3 uops)",
         "MetricGroup": "TopdownL2",
         "ScaleUnit": "1percent of slots"
-- 
2.47.0

On 13/03/2025 8:15 pm, Ilkka Koskinen wrote:
> frontend_bound metrics was miscalculated due to different scaling in
> a couple of metrics it depends on. Change the scaling to match with
> AmpereOne.
> 
> Fixes: 16438b652b46 ("perf vendor events arm64 AmpereOneX: Add core PMU events and metrics")
> Signed-off-by: Ilkka Koskinen <ilkka@os.amperecomputing.com>
> ---
>   .../arch/arm64/ampere/ampereonex/metrics.json          | 10 +++++-----
>   1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/tools/perf/pmu-events/arch/arm64/ampere/ampereonex/metrics.json b/tools/perf/pmu-events/arch/arm64/ampere/ampereonex/metrics.json
> index c5d1d22bd034..5228f94a793f 100644
> --- a/tools/perf/pmu-events/arch/arm64/ampere/ampereonex/metrics.json
> +++ b/tools/perf/pmu-events/arch/arm64/ampere/ampereonex/metrics.json
> @@ -229,19 +229,19 @@
>       },
>       {
>           "MetricName": "slots_lost_misspeculation_fraction",
> -        "MetricExpr": "(OP_SPEC - OP_RETIRED) / (CPU_CYCLES * #slots)",
> +        "MetricExpr": "100 * (OP_SPEC - OP_RETIRED) / (CPU_CYCLES * #slots)",
>           "BriefDescription": "Fraction of slots lost due to misspeculation",
>           "DefaultMetricgroupName": "TopdownL1",
>           "MetricGroup": "Default;TopdownL1",
> -        "ScaleUnit": "100percent of slots"
> +        "ScaleUnit": "1percent of slots"
>       },
>       {
>           "MetricName": "retired_fraction",
> -        "MetricExpr": "OP_RETIRED / (CPU_CYCLES * #slots)",
> +        "MetricExpr": "100 * OP_RETIRED / (CPU_CYCLES * #slots)",
>           "BriefDescription": "Fraction of slots retiring, useful work",
>           "DefaultMetricgroupName": "TopdownL1",
>           "MetricGroup": "Default;TopdownL1",
> -        "ScaleUnit": "100percent of slots"
> +        "ScaleUnit": "1percent of slots"

I never considered that ScaleUnit could be an issue when the metric was 
included in other metrics with different scales. Seems like an easy trap.

Reviewed-by: James Clark <james.clark@linaro.org>