]> git.proxmox.com Git - mirror_ubuntu-kernels.git/commitdiff
perf stat: Capitalize topdown metrics' names
authorZhengjun Xing <zhengjun.xing@linux.intel.com>
Thu, 25 Aug 2022 01:54:58 +0000 (09:54 +0800)
committerArnaldo Carvalho de Melo <acme@redhat.com>
Sat, 27 Aug 2022 14:55:17 +0000 (11:55 -0300)
Capitalize topdown metrics' names to follow the intel SDM.

Before:

 # ./perf stat -a  sleep 1

 Performance counter stats for 'system wide':

        228,094.05 msec cpu-clock                        #  225.026 CPUs utilized
               842      context-switches                 #    3.691 /sec
               224      cpu-migrations                   #    0.982 /sec
                70      page-faults                      #    0.307 /sec
        23,164,105      cycles                           #    0.000 GHz
        29,403,446      instructions                     #    1.27  insn per cycle
         5,268,185      branches                         #   23.097 K/sec
            33,239      branch-misses                    #    0.63% of all branches
       136,248,990      slots                            #  597.337 K/sec
        32,976,450      topdown-retiring                 #     24.2% retiring
         4,651,918      topdown-bad-spec                 #      3.4% bad speculation
        26,148,695      topdown-fe-bound                 #     19.2% frontend bound
        72,515,776      topdown-be-bound                 #     53.2% backend bound
         6,008,540      topdown-heavy-ops                #      4.4% heavy operations       #     19.8% light operations
         3,934,049      topdown-br-mispredict            #      2.9% branch mispredict      #      0.5% machine clears
        16,655,439      topdown-fetch-lat                #     12.2% fetch latency          #      7.0% fetch bandwidth
        41,635,972      topdown-mem-bound                #     30.5% memory bound           #     22.7% Core bound

       1.013634593 seconds time elapsed

After:

 # ./perf stat -a  sleep 1

 Performance counter stats for 'system wide':

        228,081.94 msec cpu-clock                        #  225.003 CPUs utilized
               824      context-switches                 #    3.613 /sec
               224      cpu-migrations                   #    0.982 /sec
                67      page-faults                      #    0.294 /sec
        22,647,423      cycles                           #    0.000 GHz
        28,870,551      instructions                     #    1.27  insn per cycle
         5,167,099      branches                         #   22.655 K/sec
            32,383      branch-misses                    #    0.63% of all branches
       133,411,074      slots                            #  584.926 K/sec
        32,352,607      topdown-retiring                 #     24.3% Retiring
         4,456,977      topdown-bad-spec                 #      3.3% Bad Speculation
        25,626,487      topdown-fe-bound                 #     19.2% Frontend Bound
        70,955,316      topdown-be-bound                 #     53.2% Backend Bound
         5,834,844      topdown-heavy-ops                #      4.4% Heavy Operations       #     19.9% Light Operations
         3,738,781      topdown-br-mispredict            #      2.8% Branch Mispredict      #      0.5% Machine Clears
        16,286,803      topdown-fetch-lat                #     12.2% Fetch Latency          #      7.0% Fetch Bandwidth
        40,802,069      topdown-mem-bound                #     30.6% Memory Bound           #     22.6% Core Bound

       1.013683125 seconds time elapsed

Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Xing Zhengjun <zhengjun.xing@linux.intel.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Alexander Shishkin <alexander.shishkin@intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20220825015458.3252239-1-zhengjun.xing@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
tools/perf/util/stat-shadow.c

index 979c8cb918f724f3ff71b15b1200667a7363ba38..788ce5e46470a4ad16666c20c67c5bd1e2f16722 100644 (file)
@@ -1193,7 +1193,7 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config,
                                                  &rsd);
                if (retiring > 0.7)
                        color = PERF_COLOR_GREEN;
-               print_metric(config, ctxp, color, "%8.1f%%", "retiring",
+               print_metric(config, ctxp, color, "%8.1f%%", "Retiring",
                                retiring * 100.);
        } else if (perf_stat_evsel__is(evsel, TOPDOWN_FE_BOUND) &&
                   full_td(cpu_map_idx, st, &rsd)) {
@@ -1202,7 +1202,7 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config,
                                                  &rsd);
                if (fe_bound > 0.2)
                        color = PERF_COLOR_RED;
-               print_metric(config, ctxp, color, "%8.1f%%", "frontend bound",
+               print_metric(config, ctxp, color, "%8.1f%%", "Frontend Bound",
                                fe_bound * 100.);
        } else if (perf_stat_evsel__is(evsel, TOPDOWN_BE_BOUND) &&
                   full_td(cpu_map_idx, st, &rsd)) {
@@ -1211,7 +1211,7 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config,
                                                  &rsd);
                if (be_bound > 0.2)
                        color = PERF_COLOR_RED;
-               print_metric(config, ctxp, color, "%8.1f%%", "backend bound",
+               print_metric(config, ctxp, color, "%8.1f%%", "Backend Bound",
                                be_bound * 100.);
        } else if (perf_stat_evsel__is(evsel, TOPDOWN_BAD_SPEC) &&
                   full_td(cpu_map_idx, st, &rsd)) {
@@ -1220,7 +1220,7 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config,
                                                  &rsd);
                if (bad_spec > 0.1)
                        color = PERF_COLOR_RED;
-               print_metric(config, ctxp, color, "%8.1f%%", "bad speculation",
+               print_metric(config, ctxp, color, "%8.1f%%", "Bad Speculation",
                                bad_spec * 100.);
        } else if (perf_stat_evsel__is(evsel, TOPDOWN_HEAVY_OPS) &&
                        full_td(cpu_map_idx, st, &rsd) && (config->topdown_level > 1)) {
@@ -1234,13 +1234,13 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config,
 
                if (retiring > 0.7 && heavy_ops > 0.1)
                        color = PERF_COLOR_GREEN;
-               print_metric(config, ctxp, color, "%8.1f%%", "heavy operations",
+               print_metric(config, ctxp, color, "%8.1f%%", "Heavy Operations",
                                heavy_ops * 100.);
                if (retiring > 0.7 && light_ops > 0.6)
                        color = PERF_COLOR_GREEN;
                else
                        color = NULL;
-               print_metric(config, ctxp, color, "%8.1f%%", "light operations",
+               print_metric(config, ctxp, color, "%8.1f%%", "Light Operations",
                                light_ops * 100.);
        } else if (perf_stat_evsel__is(evsel, TOPDOWN_BR_MISPREDICT) &&
                        full_td(cpu_map_idx, st, &rsd) && (config->topdown_level > 1)) {
@@ -1254,13 +1254,13 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config,
 
                if (bad_spec > 0.1 && br_mis > 0.05)
                        color = PERF_COLOR_RED;
-               print_metric(config, ctxp, color, "%8.1f%%", "branch mispredict",
+               print_metric(config, ctxp, color, "%8.1f%%", "Branch Mispredict",
                                br_mis * 100.);
                if (bad_spec > 0.1 && m_clears > 0.05)
                        color = PERF_COLOR_RED;
                else
                        color = NULL;
-               print_metric(config, ctxp, color, "%8.1f%%", "machine clears",
+               print_metric(config, ctxp, color, "%8.1f%%", "Machine Clears",
                                m_clears * 100.);
        } else if (perf_stat_evsel__is(evsel, TOPDOWN_FETCH_LAT) &&
                        full_td(cpu_map_idx, st, &rsd) && (config->topdown_level > 1)) {
@@ -1274,13 +1274,13 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config,
 
                if (fe_bound > 0.2 && fetch_lat > 0.15)
                        color = PERF_COLOR_RED;
-               print_metric(config, ctxp, color, "%8.1f%%", "fetch latency",
+               print_metric(config, ctxp, color, "%8.1f%%", "Fetch Latency",
                                fetch_lat * 100.);
                if (fe_bound > 0.2 && fetch_bw > 0.1)
                        color = PERF_COLOR_RED;
                else
                        color = NULL;
-               print_metric(config, ctxp, color, "%8.1f%%", "fetch bandwidth",
+               print_metric(config, ctxp, color, "%8.1f%%", "Fetch Bandwidth",
                                fetch_bw * 100.);
        } else if (perf_stat_evsel__is(evsel, TOPDOWN_MEM_BOUND) &&
                        full_td(cpu_map_idx, st, &rsd) && (config->topdown_level > 1)) {
@@ -1294,13 +1294,13 @@ void perf_stat__print_shadow_stats(struct perf_stat_config *config,
 
                if (be_bound > 0.2 && mem_bound > 0.2)
                        color = PERF_COLOR_RED;
-               print_metric(config, ctxp, color, "%8.1f%%", "memory bound",
+               print_metric(config, ctxp, color, "%8.1f%%", "Memory Bound",
                                mem_bound * 100.);
                if (be_bound > 0.2 && core_bound > 0.1)
                        color = PERF_COLOR_RED;
                else
                        color = NULL;
-               print_metric(config, ctxp, color, "%8.1f%%", "Core bound",
+               print_metric(config, ctxp, color, "%8.1f%%", "Core Bound",
                                core_bound * 100.);
        } else if (evsel->metric_expr) {
                generic_metric(config, evsel->metric_expr, evsel->metric_events, NULL,