[PATCH v2] perf stat: Merge uncore events by default for hybrid platform

[PATCH v2] perf stat: Merge uncore events by default for hybrid platform

[Jin Yao]

On hybrid platform, by default stat aggregates and reports the event counts
per pmu. For example,

  # perf stat -e cycles -a true

   Performance counter stats for 'system wide':

           1,400,445      cpu_core/cycles/
             680,881      cpu_atom/cycles/

         0.001770773 seconds time elapsed

While for uncore events, that's not a suitable method. Uncore has nothing
to do with hybrid. So for uncore events, we aggregate event counts from all
PMUs and report the counts without PMUs.

Before:

  # perf stat -e arb/event=0x81,umask=0x1/,arb/event=0x84,umask=0x1/ -a true

   Performance counter stats for 'system wide':

               2,058      uncore_arb_0/event=0x81,umask=0x1/
               2,028      uncore_arb_1/event=0x81,umask=0x1/
                   0      uncore_arb_0/event=0x84,umask=0x1/
                   0      uncore_arb_1/event=0x84,umask=0x1/

         0.000614498 seconds time elapsed

After:

  # perf stat -e arb/event=0x81,umask=0x1/,arb/event=0x84,umask=0x1/ -a true

   Performance counter stats for 'system wide':

               3,996      arb/event=0x81,umask=0x1/
                   0      arb/event=0x84,umask=0x1/

         0.000630046 seconds time elapsed

Of course, we also keep the '--no-merge' working for uncore events.

  # perf stat -e arb/event=0x81,umask=0x1/,arb/event=0x84,umask=0x1/ --no-merge true

   Performance counter stats for 'system wide':

               1,952      uncore_arb_0/event=0x81,umask=0x1/
               1,921      uncore_arb_1/event=0x81,umask=0x1/
                   0      uncore_arb_0/event=0x84,umask=0x1/
                   0      uncore_arb_1/event=0x84,umask=0x1/

         0.000575536 seconds time elapsed

Signed-off-by: Jin Yao <yao.jin@linux.intel.com>
---
v2:
 - Use evsel__find_pmu() to find uncore pmu.
 - Create hybrid_uniquify() to check if uniquify the event name for hybrid.

 tools/perf/builtin-stat.c      |  3 ---
 tools/perf/util/stat-display.c | 14 +++++++++++++-
 2 files changed, 13 insertions(+), 4 deletions(-)

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index f9f74a514315..b67a44982b61 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -2442,9 +2442,6 @@ int cmd_stat(int argc, const char **argv)
 
 	evlist__check_cpu_maps(evsel_list);
 
-	if (perf_pmu__has_hybrid())
-		stat_config.no_merge = true;
-
 	/*
 	 * Initialize thread_map with comm names,
 	 * so we could print it out on output.
diff --git a/tools/perf/util/stat-display.c b/tools/perf/util/stat-display.c
index c588a6b7a8db..87f77016b9cc 100644
--- a/tools/perf/util/stat-display.c
+++ b/tools/perf/util/stat-display.c
@@ -593,6 +593,18 @@ static void collect_all_aliases(struct perf_stat_config *config, struct evsel *c
 	}
 }
 
+static bool is_uncore(struct evsel *evsel)
+{
+	struct perf_pmu *pmu = evsel__find_pmu(evsel);
+
+	return pmu && pmu->is_uncore;
+}
+
+static bool hybrid_uniquify(struct evsel *evsel)
+{
+	return perf_pmu__has_hybrid() && !is_uncore(evsel);
+}
+
 static bool collect_data(struct perf_stat_config *config, struct evsel *counter,
 			    void (*cb)(struct perf_stat_config *config, struct evsel *counter, void *data,
 				       bool first),
@@ -601,7 +613,7 @@ static bool collect_data(struct perf_stat_config *config, struct evsel *counter,
 	if (counter->merged_stat)
 		return false;
 	cb(config, counter, data, true);
-	if (config->no_merge)
+	if (config->no_merge || hybrid_uniquify(counter))
 		uniquify_event_name(counter);
 	else if (counter->auto_merge_stats)
 		collect_all_aliases(config, counter, cb, data);
-- 
2.27.0

Re: [PATCH v2] perf stat: Merge uncore events by default for hybrid platform

[Jiri Olsa]

On Wed, Jul 07, 2021 at 01:56:52PM +0800, Jin Yao wrote:
> On hybrid platform, by default stat aggregates and reports the event counts
> per pmu. For example,
> 
>   # perf stat -e cycles -a true
> 
>    Performance counter stats for 'system wide':
> 
>            1,400,445      cpu_core/cycles/
>              680,881      cpu_atom/cycles/
> 
>          0.001770773 seconds time elapsed
> 
> While for uncore events, that's not a suitable method. Uncore has nothing
> to do with hybrid. So for uncore events, we aggregate event counts from all
> PMUs and report the counts without PMUs.
> 
> Before:
> 
>   # perf stat -e arb/event=0x81,umask=0x1/,arb/event=0x84,umask=0x1/ -a true
> 
>    Performance counter stats for 'system wide':
> 
>                2,058      uncore_arb_0/event=0x81,umask=0x1/
>                2,028      uncore_arb_1/event=0x81,umask=0x1/
>                    0      uncore_arb_0/event=0x84,umask=0x1/
>                    0      uncore_arb_1/event=0x84,umask=0x1/
> 
>          0.000614498 seconds time elapsed
> 
> After:
> 
>   # perf stat -e arb/event=0x81,umask=0x1/,arb/event=0x84,umask=0x1/ -a true
> 
>    Performance counter stats for 'system wide':
> 
>                3,996      arb/event=0x81,umask=0x1/
>                    0      arb/event=0x84,umask=0x1/
> 
>          0.000630046 seconds time elapsed
> 
> Of course, we also keep the '--no-merge' working for uncore events.
> 
>   # perf stat -e arb/event=0x81,umask=0x1/,arb/event=0x84,umask=0x1/ --no-merge true
> 
>    Performance counter stats for 'system wide':
> 
>                1,952      uncore_arb_0/event=0x81,umask=0x1/
>                1,921      uncore_arb_1/event=0x81,umask=0x1/
>                    0      uncore_arb_0/event=0x84,umask=0x1/
>                    0      uncore_arb_1/event=0x84,umask=0x1/
> 
>          0.000575536 seconds time elapsed
> 
> Signed-off-by: Jin Yao <yao.jin@linux.intel.com>
> ---
> v2:
>  - Use evsel__find_pmu() to find uncore pmu.
>  - Create hybrid_uniquify() to check if uniquify the event name for hybrid.

Acked-by: Jiri Olsa <jolsa@redhat.com>

thanks,
jirka


> 
>  tools/perf/builtin-stat.c      |  3 ---
>  tools/perf/util/stat-display.c | 14 +++++++++++++-
>  2 files changed, 13 insertions(+), 4 deletions(-)
> 
> diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
> index f9f74a514315..b67a44982b61 100644
> --- a/tools/perf/builtin-stat.c
> +++ b/tools/perf/builtin-stat.c
> @@ -2442,9 +2442,6 @@ int cmd_stat(int argc, const char **argv)
>  
>  	evlist__check_cpu_maps(evsel_list);
>  
> -	if (perf_pmu__has_hybrid())
> -		stat_config.no_merge = true;
> -
>  	/*
>  	 * Initialize thread_map with comm names,
>  	 * so we could print it out on output.
> diff --git a/tools/perf/util/stat-display.c b/tools/perf/util/stat-display.c
> index c588a6b7a8db..87f77016b9cc 100644
> --- a/tools/perf/util/stat-display.c
> +++ b/tools/perf/util/stat-display.c
> @@ -593,6 +593,18 @@ static void collect_all_aliases(struct perf_stat_config *config, struct evsel *c
>  	}
>  }
>  
> +static bool is_uncore(struct evsel *evsel)
> +{
> +	struct perf_pmu *pmu = evsel__find_pmu(evsel);
> +
> +	return pmu && pmu->is_uncore;
> +}
> +
> +static bool hybrid_uniquify(struct evsel *evsel)
> +{
> +	return perf_pmu__has_hybrid() && !is_uncore(evsel);
> +}
> +
>  static bool collect_data(struct perf_stat_config *config, struct evsel *counter,
>  			    void (*cb)(struct perf_stat_config *config, struct evsel *counter, void *data,
>  				       bool first),
> @@ -601,7 +613,7 @@ static bool collect_data(struct perf_stat_config *config, struct evsel *counter,
>  	if (counter->merged_stat)
>  		return false;
>  	cb(config, counter, data, true);
> -	if (config->no_merge)
> +	if (config->no_merge || hybrid_uniquify(counter))
>  		uniquify_event_name(counter);
>  	else if (counter->auto_merge_stats)
>  		collect_all_aliases(config, counter, cb, data);
> -- 
> 2.27.0
> 

Re: [PATCH v2] perf stat: Merge uncore events by default for hybrid platform

[Arnaldo Carvalho de Melo]

Em Sun, Jul 11, 2021 at 06:02:13PM +0200, Jiri Olsa escreveu:
> On Wed, Jul 07, 2021 at 01:56:52PM +0800, Jin Yao wrote:
<SNIP>
> > While for uncore events, that's not a suitable method. Uncore has nothing
> > to do with hybrid. So for uncore events, we aggregate event counts from all
> > PMUs and report the counts without PMUs.
<SNIP>
 
> Acked-by: Jiri Olsa <jolsa@redhat.com>

Thanks, applied to perf/urgent.

- Arnaldo