Re: [PATCH] perf vendor events: Update Skylake client events to v50
> You mean change event-converter-for-linux-perf to add this as JSON > comments at the start of the generated files? JSON doesn't support comments unfortunately (unless they're somehow part of the data schema) -Andi
Re: [PATCH] perf vendor events: Update Skylake client events to v50
Em Mon, Nov 16, 2020 at 11:12:30AM -0800, Ian Rogers escreveu: > On Mon, Nov 16, 2020 at 9:05 AM Arnaldo Carvalho de Melo > wrote: > > > Em Fri, Nov 06, 2020 at 01:21:58PM +0900, Namhyung Kim escreveu: > > > On Fri, Nov 6, 2020 at 12:12 PM Jin, Yao > > wrote: > > > > >> Signed-off-by: Jin Yao > > > > > > > > > > It seems not applied to acme/perf/core cleanly. > > > > > > > > > > Thanks, > > > > > Namhyung > > > > > > > > > > > > > It seems the patch mail is truncated. :( > > > > > > > > I attach the patch file in the mail. Sorry for inconvenience. > > > > > > I've checked it fixed the perf test on my laptop (skylake). > > > > > > Tested-by: Namhyung Kim > > > > Thanks, applied. > > > > - Arnaldo > > > > Nit, as the code is generated, would it be possible to add the commands to > regenerate it? > > E.g. Using the code in: > https://github.com/intel/event-converter-for-linux-perf > And extracted perfmon_server_events_v1.4.tar from: > https://download.01.org/perfmon/ > run: > json-to-perf-json.py --outdir tools/perf/pmu-events/arch/arch/x86/sklylakex > perfmon/SKX/skylakex_core_v1.24.json > Looking at the download.01.org/perfmon json there are files > like skylakex_fp_arith_inst_v1.24.json, and how these are incorporated into > these events is less than clear. You mean change event-converter-for-linux-perf to add this as JSON comments at the start of the generated files? - Arnaldo
Re: [PATCH] perf vendor events: Update Skylake client events to v50
Em Fri, Nov 06, 2020 at 01:21:58PM +0900, Namhyung Kim escreveu: > On Fri, Nov 6, 2020 at 12:12 PM Jin, Yao wrote: > > >> Signed-off-by: Jin Yao > > > > > > It seems not applied to acme/perf/core cleanly. > > > > > > Thanks, > > > Namhyung > > > > > > > It seems the patch mail is truncated. :( > > > > I attach the patch file in the mail. Sorry for inconvenience. > > I've checked it fixed the perf test on my laptop (skylake). > > Tested-by: Namhyung Kim Thanks, applied. - Arnaldo
Re: [PATCH] perf vendor events: Update Skylake client events to v50
>I'd prefer if we could make a copy of these scripts into the kernel tree >along with the data files from: >https://download.01.org/perfmon/ FWIW I originally tried this to include the raw JSON files, but Ingo objected and wanted the files to be split up[1] and avoiding redundant headers. That's why we ended up with the "compiled" format. [1] https://lkml.org/lkml/2015/5/28/336 >Having a generated file that can't be edited, reviewed, .. I also worry >that if/when we change the json format then event-converter-for-linux-perf >will need to support a multitude of perf versions. You mean for backports? The assumption was that perf backports would backport any perf changes for new formats too. In general I generally discourage any perf tools backport, people should just use a new version of the perf tool on old kernels. In principle the scripts could be included, but without the raw files it would be somewhat pointless. -andi
Re: [PATCH] perf vendor events: Update Skylake client events to v50
Hi Ian, On 11/16/2020 3:25 PM, Ian Rogers wrote: No given metric group: L2_Evictions_Silent_PKI L2_Evictions_NonSilent_PKI These two metrics are available on SKX/CLX. From TMA_Metrics.csv, they are under the column 'SKX' not under the column 'SKL/KBL'. Thanks Jin Yao
Re: [PATCH] perf vendor events: Update Skylake client events to v50
> Can I get ACK for this patch? > > It's mainly for fixing the perf-test issue and MEM_Parallel_Reads issue. Acked-by: Andi Kleen The JSON updates should be imho just merged automatically. Not much point in doing code review on them. If there's a problem it has to be fixed JSON upstream anyways. -Andi
Re: [PATCH] perf vendor events: Update Skylake client events to v50
Hi Andi, Ian, On 11/6/2020 12:21 PM, Namhyung Kim wrote: On Fri, Nov 6, 2020 at 12:12 PM Jin, Yao wrote: Hi Namhyung, On 11/6/2020 11:00 AM, Namhyung Kim wrote: Hi, On Wed, Nov 04, 2020 at 09:51:02AM +0800, Jin Yao wrote: - Update Skylake events to v50. - Update Skylake JSON metrics from TMAM 4.0. - Fix the issue in DRAM_Parallel_Reads - Fix the perf test warning Before: root@kbl-ppc:~# perf stat -M DRAM_Parallel_Reads -- sleep 1 event syntax error: '{arb/event=0x80,umask=0x2/,arb/event=0x80,umask=0x2,thresh=1/}:W' \___ unknown term 'thresh' for pmu 'uncore_arb' valid terms: event,edge,inv,umask,cmask,config,config1,config2,name,period,percore Initial error: event syntax error: '..umask=0x2/,arb/event=0x80,umask=0x2,thresh=1/}:W' \___ Cannot find PMU `arb'. Missing kernel support? root@kbl-ppc:~# perf test metrics 10: PMU events : 10.3: Parsing of PMU event table metrics : Skip (some metrics failed) 10.4: Parsing of PMU event table metrics with fake PMUs: Ok 67: Parse and process metrics : Ok After: root@kbl-ppc:~# perf stat -M MEM_Parallel_Reads -- sleep 1 Performance counter stats for 'system wide': 4,951,646 arb/event=0x80,umask=0x2/ #26.30 MEM_Parallel_Reads (50.04%) 188,251 arb/event=0x80,umask=0x2,cmask=1/ (49.96%) 1.000867010 seconds time elapsed root@kbl-ppc:~# perf test metrics 10: PMU events : 10.3: Parsing of PMU event table metrics : Ok 10.4: Parsing of PMU event table metrics with fake PMUs: Ok 67: Parse and process metrics : Ok Signed-off-by: Jin Yao It seems not applied to acme/perf/core cleanly. Thanks, Namhyung It seems the patch mail is truncated. :( I attach the patch file in the mail. Sorry for inconvenience. I've checked it fixed the perf test on my laptop (skylake). Tested-by: Namhyung Kim Thanks, Namhyung Can I get ACK for this patch? It's mainly for fixing the perf-test issue and MEM_Parallel_Reads issue. Thanks Jin Yao
Re: [PATCH] perf vendor events: Update Skylake client events to v50
On Fri, Nov 6, 2020 at 12:12 PM Jin, Yao wrote: > > Hi Namhyung, > > On 11/6/2020 11:00 AM, Namhyung Kim wrote: > > Hi, > > > > On Wed, Nov 04, 2020 at 09:51:02AM +0800, Jin Yao wrote: > >> - Update Skylake events to v50. > >> - Update Skylake JSON metrics from TMAM 4.0. > >> - Fix the issue in DRAM_Parallel_Reads > >> - Fix the perf test warning > >> > >> Before: > >> > >> root@kbl-ppc:~# perf stat -M DRAM_Parallel_Reads -- sleep 1 > >> event syntax error: > >> '{arb/event=0x80,umask=0x2/,arb/event=0x80,umask=0x2,thresh=1/}:W' > >> \___ unknown term 'thresh' for pmu 'uncore_arb' > >> > >> valid terms: > >> event,edge,inv,umask,cmask,config,config1,config2,name,period,percore > >> > >> Initial error: > >> event syntax error: '..umask=0x2/,arb/event=0x80,umask=0x2,thresh=1/}:W' > >>\___ Cannot find PMU `arb'. Missing > >> kernel support? > >> > >> root@kbl-ppc:~# perf test metrics > >> 10: PMU events : > >> 10.3: Parsing of PMU event table metrics : Skip (some > >> metrics failed) > >> 10.4: Parsing of PMU event table metrics with fake PMUs: Ok > >> 67: Parse and process metrics : Ok > >> > >> After: > >> > >> root@kbl-ppc:~# perf stat -M MEM_Parallel_Reads -- sleep 1 > >> > >> Performance counter stats for 'system wide': > >> > >> 4,951,646 arb/event=0x80,umask=0x2/ #26.30 > >> MEM_Parallel_Reads (50.04%) > >> 188,251 arb/event=0x80,umask=0x2,cmask=1/ > >> (49.96%) > >> > >> 1.000867010 seconds time elapsed > >> > >> root@kbl-ppc:~# perf test metrics > >> 10: PMU events : > >> 10.3: Parsing of PMU event table metrics : Ok > >> 10.4: Parsing of PMU event table metrics with fake PMUs: Ok > >> 67: Parse and process metrics : Ok > >> > >> Signed-off-by: Jin Yao > > > > It seems not applied to acme/perf/core cleanly. > > > > Thanks, > > Namhyung > > > > It seems the patch mail is truncated. :( > > I attach the patch file in the mail. Sorry for inconvenience. I've checked it fixed the perf test on my laptop (skylake). Tested-by: Namhyung Kim Thanks, Namhyung
Re: [PATCH] perf vendor events: Update Skylake client events to v50
Hi, On Wed, Nov 04, 2020 at 09:51:02AM +0800, Jin Yao wrote: > - Update Skylake events to v50. > - Update Skylake JSON metrics from TMAM 4.0. > - Fix the issue in DRAM_Parallel_Reads > - Fix the perf test warning > > Before: > > root@kbl-ppc:~# perf stat -M DRAM_Parallel_Reads -- sleep 1 > event syntax error: > '{arb/event=0x80,umask=0x2/,arb/event=0x80,umask=0x2,thresh=1/}:W' > \___ unknown term 'thresh' for pmu 'uncore_arb' > > valid terms: > event,edge,inv,umask,cmask,config,config1,config2,name,period,percore > > Initial error: > event syntax error: '..umask=0x2/,arb/event=0x80,umask=0x2,thresh=1/}:W' > \___ Cannot find PMU `arb'. Missing kernel > support? > > root@kbl-ppc:~# perf test metrics > 10: PMU events : > 10.3: Parsing of PMU event table metrics : Skip (some metrics > failed) > 10.4: Parsing of PMU event table metrics with fake PMUs: Ok > 67: Parse and process metrics : Ok > > After: > > root@kbl-ppc:~# perf stat -M MEM_Parallel_Reads -- sleep 1 > > Performance counter stats for 'system wide': > > 4,951,646 arb/event=0x80,umask=0x2/ #26.30 > MEM_Parallel_Reads (50.04%) >188,251 arb/event=0x80,umask=0x2,cmask=1/ > (49.96%) > >1.000867010 seconds time elapsed > > root@kbl-ppc:~# perf test metrics > 10: PMU events : > 10.3: Parsing of PMU event table metrics : Ok > 10.4: Parsing of PMU event table metrics with fake PMUs: Ok > 67: Parse and process metrics : Ok > > Signed-off-by: Jin Yao It seems not applied to acme/perf/core cleanly. Thanks, Namhyung