cpufreq: intel_pstate: Update Balance performance EPP for Emerald Rapids

[PATCH] cpufreq: intel_pstate: Update Balance performance EPP for Emerald Rapids

Posted by Pedro Henrique Kopper 1 year, 4 months ago

On Intel Emerald Rapids machines, we ship the Energy Performance Preference
(EPP) default for balance_performance as 128. However, during an internal
investigation together with Intel, we have determined that 32 is a more
suitable value. This leads to significant improvements in both performance
and energy:

POV-Ray: 32% faster | 12% less energy
OpenSSL: 12% faster | energy within 1%
Build Linux Kernel: 29% faster | 18% less energy

Therefore, we should move the default EPP for balance_performance to 32.
This is in line with what has already been done for Sapphire Rapids.

Signed-off-by: Pedro Henrique Kopper <pedro.kopper@canonical.com>
---
 drivers/cpufreq/intel_pstate.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c
index 392a8000b238..c0278d023cfc 100644
--- a/drivers/cpufreq/intel_pstate.c
+++ b/drivers/cpufreq/intel_pstate.c
@@ -3405,6 +3405,7 @@ static const struct x86_cpu_id intel_epp_default[] = {
 	 */
 	X86_MATCH_VFM(INTEL_ALDERLAKE_L, HWP_SET_DEF_BALANCE_PERF_EPP(102)),
 	X86_MATCH_VFM(INTEL_SAPPHIRERAPIDS_X, HWP_SET_DEF_BALANCE_PERF_EPP(32)),
+	X86_MATCH_VFM(INTEL_EMERALDRAPIDS_X, HWP_SET_DEF_BALANCE_PERF_EPP(32)),
 	X86_MATCH_VFM(INTEL_METEORLAKE_L, HWP_SET_EPP_VALUES(HWP_EPP_POWERSAVE,
 		      179, 64, 16)),
 	X86_MATCH_VFM(INTEL_ARROWLAKE, HWP_SET_EPP_VALUES(HWP_EPP_POWERSAVE,
-- 
2.43.0

Re: [PATCH] cpufreq: intel_pstate: Update Balance performance EPP for Emerald Rapids

Posted by srinivas pandruvada 1 year, 4 months ago

On Thu, 2024-08-01 at 13:41 -0300, Pedro Henrique Kopper wrote:
> On Intel Emerald Rapids machines, we ship the Energy Performance
> Preference
> (EPP) default for balance_performance as 128. However, during an
> internal
> investigation together with Intel, we have determined that 32 is a
> more
> suitable value. This leads to significant improvements in both
> performance
> and energy:
> 
> POV-Ray: 32% faster | 12% less energy
> OpenSSL: 12% faster | energy within 1%
> Build Linux Kernel: 29% faster | 18% less energy
> 
> Therefore, we should move the default EPP for balance_performance to
> 32.
> This is in line with what has already been done for Sapphire Rapids.
> 
> Signed-off-by: Pedro Henrique Kopper <pedro.kopper@canonical.com>
Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>

> ---
>  drivers/cpufreq/intel_pstate.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/cpufreq/intel_pstate.c
> b/drivers/cpufreq/intel_pstate.c
> index 392a8000b238..c0278d023cfc 100644
> --- a/drivers/cpufreq/intel_pstate.c
> +++ b/drivers/cpufreq/intel_pstate.c
> @@ -3405,6 +3405,7 @@ static const struct x86_cpu_id
> intel_epp_default[] = {
>  	 */
>  	X86_MATCH_VFM(INTEL_ALDERLAKE_L,
> HWP_SET_DEF_BALANCE_PERF_EPP(102)),
>  	X86_MATCH_VFM(INTEL_SAPPHIRERAPIDS_X,
> HWP_SET_DEF_BALANCE_PERF_EPP(32)),
> +	X86_MATCH_VFM(INTEL_EMERALDRAPIDS_X,
> HWP_SET_DEF_BALANCE_PERF_EPP(32)),
>  	X86_MATCH_VFM(INTEL_METEORLAKE_L,
> HWP_SET_EPP_VALUES(HWP_EPP_POWERSAVE,
>  		      179, 64, 16)),
>  	X86_MATCH_VFM(INTEL_ARROWLAKE,
> HWP_SET_EPP_VALUES(HWP_EPP_POWERSAVE,

Re: [PATCH] cpufreq: intel_pstate: Update Balance performance EPP for Emerald Rapids

Posted by Rafael J. Wysocki 1 year, 4 months ago

On Thu, Aug 1, 2024 at 9:42 PM srinivas pandruvada
<srinivas.pandruvada@linux.intel.com> wrote:
>
> On Thu, 2024-08-01 at 13:41 -0300, Pedro Henrique Kopper wrote:
> > On Intel Emerald Rapids machines, we ship the Energy Performance
> > Preference
> > (EPP) default for balance_performance as 128. However, during an
> > internal
> > investigation together with Intel, we have determined that 32 is a
> > more
> > suitable value. This leads to significant improvements in both
> > performance
> > and energy:
> >
> > POV-Ray: 32% faster | 12% less energy
> > OpenSSL: 12% faster | energy within 1%
> > Build Linux Kernel: 29% faster | 18% less energy
> >
> > Therefore, we should move the default EPP for balance_performance to
> > 32.
> > This is in line with what has already been done for Sapphire Rapids.
> >
> > Signed-off-by: Pedro Henrique Kopper <pedro.kopper@canonical.com>
> Acked-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>

Applied as 6.11-rc material.

If I'm able to send a pull request next week, I'll push this for -rc3.
Otherwise, it'll be -rc4.

> > ---
> >  drivers/cpufreq/intel_pstate.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/cpufreq/intel_pstate.c
> > b/drivers/cpufreq/intel_pstate.c
> > index 392a8000b238..c0278d023cfc 100644
> > --- a/drivers/cpufreq/intel_pstate.c
> > +++ b/drivers/cpufreq/intel_pstate.c
> > @@ -3405,6 +3405,7 @@ static const struct x86_cpu_id
> > intel_epp_default[] = {
> >        */
> >       X86_MATCH_VFM(INTEL_ALDERLAKE_L,
> > HWP_SET_DEF_BALANCE_PERF_EPP(102)),
> >       X86_MATCH_VFM(INTEL_SAPPHIRERAPIDS_X,
> > HWP_SET_DEF_BALANCE_PERF_EPP(32)),
> > +     X86_MATCH_VFM(INTEL_EMERALDRAPIDS_X,
> > HWP_SET_DEF_BALANCE_PERF_EPP(32)),
> >       X86_MATCH_VFM(INTEL_METEORLAKE_L,
> > HWP_SET_EPP_VALUES(HWP_EPP_POWERSAVE,
> >                     179, 64, 16)),
> >       X86_MATCH_VFM(INTEL_ARROWLAKE,
> > HWP_SET_EPP_VALUES(HWP_EPP_POWERSAVE,
>
>