[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAP-5=fWcBPz6XM0PeRaQk+_fKFzqrvL_ESA4g9p6-sjysSrCyA@mail.gmail.com>
Date: Thu, 6 Apr 2023 17:06:33 -0700
From: Ian Rogers <irogers@...gle.com>
To: "Liang, Kan" <kan.liang@...ux.intel.com>
Cc: Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Jiri Olsa <jolsa@...nel.org>,
Namhyung Kim <namhyung@...nel.org>,
Adrian Hunter <adrian.hunter@...el.com>,
Zhengjun Xing <zhengjun.xing@...ux.intel.com>,
linux-kernel@...r.kernel.org, linux-perf-users@...r.kernel.org
Subject: Re: [PATCH v1 1/5] perf vendor events intel: Update free running
alderlake events
On Thu, Apr 6, 2023 at 1:13 PM Liang, Kan <kan.liang@...ux.intel.com> wrote:
>
>
>
> On 2023-04-06 2:46 p.m., Ian Rogers wrote:
> > Fix the PMU name, event code and umask.
> >
> > These updates were generated by:
> > https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
> > with this PR:
> > https://github.com/intel/perfmon/pull/66
> >
> > Signed-off-by: Ian Rogers <irogers@...gle.com>
> > ---
> > .../arch/x86/alderlake/uncore-memory.json | 16 ++++++++++++----
> > .../arch/x86/alderlaken/uncore-memory.json | 16 ++++++++++++----
> > 2 files changed, 24 insertions(+), 8 deletions(-)
> >
> > diff --git a/tools/perf/pmu-events/arch/x86/alderlake/uncore-memory.json b/tools/perf/pmu-events/arch/x86/alderlake/uncore-memory.json
> > index 2ccd9cf96957..ea25bb411f89 100644
> > --- a/tools/perf/pmu-events/arch/x86/alderlake/uncore-memory.json
> > +++ b/tools/perf/pmu-events/arch/x86/alderlake/uncore-memory.json
> > @@ -1,29 +1,37 @@
> > [
> > {
> > "BriefDescription": "Counts every 64B read request entering the Memory Controller 0 to DRAM (sum of all channels).",
> > + "EventCode": "0xff",
> > "EventName": "UNC_MC0_RDCAS_COUNT_FREERUN",
> > "PerPkg": "1",
> > "PublicDescription": "Counts every 64B read request entering the Memory Controller 0 to DRAM (sum of all channels).",
> > - "Unit": "iMC"
> > + "UMask": "0x20",
> > + "Unit": "imc_free_running"
>
> The imc_free_running will give the counter value for both imc0 and imc1.
>
> I think we need "Unit": "imc_free_running_0" here. But I'm not sure if
> the perf tool can handle it.
Thanks Kan, it can handle it so I'll update in v2.
Ian
> Thanks,
> Kan
>
> > },
> > {
> > "BriefDescription": "Counts every 64B write request entering the Memory Controller 0 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
> > + "EventCode": "0xff",
> > "EventName": "UNC_MC0_WRCAS_COUNT_FREERUN",
> > "PerPkg": "1",
> > - "Unit": "iMC"
> > + "UMask": "0x30",
> > + "Unit": "imc_free_running"
> > },
> > {
> > "BriefDescription": "Counts every 64B read request entering the Memory Controller 1 to DRAM (sum of all channels).",
> > + "EventCode": "0xff",
> > "EventName": "UNC_MC1_RDCAS_COUNT_FREERUN",
> > "PerPkg": "1",
> > "PublicDescription": "Counts every 64B read entering the Memory Controller 1 to DRAM (sum of all channels).",
> > - "Unit": "iMC"
> > + "UMask": "0x20",
> > + "Unit": "imc_free_running"
> > },
> > {
> > "BriefDescription": "Counts every 64B write request entering the Memory Controller 1 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
> > + "EventCode": "0xff",
> > "EventName": "UNC_MC1_WRCAS_COUNT_FREERUN",
> > "PerPkg": "1",
> > - "Unit": "iMC"
> > + "UMask": "0x30",
> > + "Unit": "imc_free_running"
> > },
> > {
> > "BriefDescription": "ACT command for a read request sent to DRAM",
> > diff --git a/tools/perf/pmu-events/arch/x86/alderlaken/uncore-memory.json b/tools/perf/pmu-events/arch/x86/alderlaken/uncore-memory.json
> > index 2ccd9cf96957..ea25bb411f89 100644
> > --- a/tools/perf/pmu-events/arch/x86/alderlaken/uncore-memory.json
> > +++ b/tools/perf/pmu-events/arch/x86/alderlaken/uncore-memory.json
> > @@ -1,29 +1,37 @@
> > [
> > {
> > "BriefDescription": "Counts every 64B read request entering the Memory Controller 0 to DRAM (sum of all channels).",
> > + "EventCode": "0xff",
> > "EventName": "UNC_MC0_RDCAS_COUNT_FREERUN",
> > "PerPkg": "1",
> > "PublicDescription": "Counts every 64B read request entering the Memory Controller 0 to DRAM (sum of all channels).",
> > - "Unit": "iMC"
> > + "UMask": "0x20",
> > + "Unit": "imc_free_running"
> > },
> > {
> > "BriefDescription": "Counts every 64B write request entering the Memory Controller 0 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
> > + "EventCode": "0xff",
> > "EventName": "UNC_MC0_WRCAS_COUNT_FREERUN",
> > "PerPkg": "1",
> > - "Unit": "iMC"
> > + "UMask": "0x30",
> > + "Unit": "imc_free_running"
> > },
> > {
> > "BriefDescription": "Counts every 64B read request entering the Memory Controller 1 to DRAM (sum of all channels).",
> > + "EventCode": "0xff",
> > "EventName": "UNC_MC1_RDCAS_COUNT_FREERUN",
> > "PerPkg": "1",
> > "PublicDescription": "Counts every 64B read entering the Memory Controller 1 to DRAM (sum of all channels).",
> > - "Unit": "iMC"
> > + "UMask": "0x20",
> > + "Unit": "imc_free_running"
> > },
> > {
> > "BriefDescription": "Counts every 64B write request entering the Memory Controller 1 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
> > + "EventCode": "0xff",
> > "EventName": "UNC_MC1_WRCAS_COUNT_FREERUN",
> > "PerPkg": "1",
> > - "Unit": "iMC"
> > + "UMask": "0x30",
> > + "Unit": "imc_free_running"
> > },
> > {
> > "BriefDescription": "ACT command for a read request sent to DRAM",
Powered by blists - more mailing lists