perf stat: Merge uncore events by default for hybrid platform
PATCH v2 1/1 perf stat: Merge uncore events by default for hybrid platform
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 elapsedWhile 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 elapsedOf 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
---
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(-)
tools/perf/builtin-stat.c +0 -3
@@ -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.
tools/perf/util/stat-display.c +13 -1
@@ -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);
Signature · 1 line
2.27.0
Quoted 53 lines · Jin Yao
On hybrid platform, by default stat aggregates and reports the event counts per pmu. For example,
# perf stat -e cycles -a truePerformance counter stats for 'system wide': 1,400,445 cpu_core/cycles/ 680,881 cpu_atom/cycles/ 0.001770773 seconds time elapsedWhile 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 truePerformance 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 elapsedAfter:
# perf stat -e arb/event=0x81,umask=0x1/,arb/event=0x84,umask=0x1/ -a truePerformance counter stats for 'system wide': 3,996 arb/event=0x81,umask=0x1/ 0 arb/event=0x84,umask=0x1/ 0.000630046 seconds time elapsedOf 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 truePerformance 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
Quoted 5 lines
Signed-off-by Jin Yao <[email protected]>---
v2:
- Useevsel__find_pmu()to find uncore pmu.
- Createhybrid_uniquify()to check if uniquify the event name for hybrid.
thanks,
jirka
Quoted 54 lines
tools/perf/builtin-stat.c | 3 --- tools/perf/util/stat-display.c | 14 +++++++++++++- 2 files changed, 13 insertions(+), 4 deletions(-)tools/perf/builtin-stat.c +0
-3@@ -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. tools/perf/util/stat-display.c +13
-1@@ -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);Signature · 1 line
2.27.0
Em Sun, Jul 11, 2021 at 06:02:13PM +0200, Jiri Olsa escreveu:
Quoted 1 line · Jin Yao
<SNIP>
Quoted 3 lines
Quoted 3 lines
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>
Quoted 1 line
Acked-by Jiri Olsa <[email protected]>
Thanks, applied to perf/urgent.
- Arnaldo