Skip to contentSkip to conversation

perf stat: Add Topdown metrics L2 events as default events

PATCH Applied [email protected] 3 participants 3 messages Jul 8, 2021 – Jul 9, 2021 perf events

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 perf stat: Add Topdown metrics L2 events as default events

From: Kan Liang <[email protected]>

The Topdown Microarchitecture Analysis (TMA) Method is a structured analysis methodology to identify critical performance bottlenecks in out-of-order processors. The Topdown metrics L1 event was added as default events in the commit 42641d6f4d15 ("perf stat: Add Topdown metrics events as default events"). From the Sapphire Rapids server and later platforms, the same dedicated "metrics" register is extended to support both L1 and L2 events. Add both L1 and L2 Topdown metrics events as default events to enrich the default measuring information, if the new measurement register is available. On legacy systems there is no change to avoid extra multiplexing.

The topdown_level indicates the max metrics level for the top-down statistics. Set it to 2 to display all L1 and L2 Topdown metrics events.

With the patch:

 $ perf stat sleep 1

Performance counter stats for 'sleep 1':

              0.59 msec task-clock                #    0.001 CPUs utilized
                 1      context-switches          #    1.687 K/sec
                 0      cpu-migrations            #    0.000 /sec
                76      page-faults               #  128.198 K/sec
         1,405,318      cycles                    #    2.371 GHz
         1,471,136      instructions              #    1.05  insn per cycle
           310,132      branches                  #  523.136 M/sec
            10,435      branch-misses             #    3.36% of all branches
         8,431,908      slots                     #   14.223 G/sec
         1,554,116      topdown-retiring          #     18.4% retiring
         1,289,585      topdown-bad-spec          #     15.2% bad speculation
         2,810,636      topdown-fe-bound          #     33.2% frontend bound
         2,810,636      topdown-be-bound          #     33.2% backend bound
           231,464      topdown-heavy-ops         #      2.7% heavy operations       #     15.6% light operations
         1,223,453      topdown-br-mispredict     #     14.5% branch mispredict      #      0.8% machine clears
         1,884,779      topdown-fetch-lat         #     22.3% fetch latency          #     10.9% fetch bandwidth
         1,454,917      topdown-mem-bound         #     17.2% memory bound           #     16.0% Core bound
       1.001179699 seconds time elapsed

       0.000000000 seconds user
       0.001238000 seconds sys
---
 tools/perf/arch/x86/util/evlist.c | 6 +++++-
 tools/perf/builtin-stat.c         | 3 +++
 2 files changed, 8 insertions(+), 1 deletion(-)
tools/perf/arch/x86/util/evlist.c +5 -1
index 8c6732c..0b09510 100644
@@ -5,11 +5,15 @@
 #include "util/parse-events.h"
 
 #define TOPDOWN_L1_EVENTS	"{slots,topdown-retiring,topdown-bad-spec,topdown-fe-bound,topdown-be-bound}"
+#define TOPDOWN_L2_EVENTS	"{slots,topdown-retiring,topdown-bad-spec,topdown-fe-bound,topdown-be-bound,topdown-heavy-ops,topdown-br-mispredict,topdown-fetch-lat,topdown-mem-bound}"
 
 int arch_evlist__add_default_attrs(struct evlist *evlist)
 {
 	if (!pmu_have_event("cpu", "slots"))
 		return 0;
 
-	return parse_events(evlist, TOPDOWN_L1_EVENTS, NULL);
+	if (pmu_have_event("cpu", "topdown-heavy-ops"))
+		return parse_events(evlist, TOPDOWN_L2_EVENTS, NULL);
+	else
+		return parse_events(evlist, TOPDOWN_L1_EVENTS, NULL);
 }
tools/perf/builtin-stat.c +3 -0
index 5a830ae..a5e93d4 100644
@@ -154,6 +154,8 @@ static const char *topdown_metric_L2_attrs[] = {
 	NULL,
 };
 
+#define TOPDOWN_MAX_LEVEL			2
+
 static const char *smi_cost_attrs = {
 	"{"
 	"msr/aperf/,"
@@ -1924,6 +1926,7 @@ static int add_default_attributes(void)
 		if (evlist__add_default_attrs(evsel_list, default_attrs1) < 0)
 			return -1;
 
+		stat_config.topdown_level = TOPDOWN_MAX_LEVEL;
 		if (arch_evlist__add_default_attrs(evsel_list) < 0)
 			return -1;
 	}
Signature · 1 line
2.7.4
maintainer Acked-by on v1 1/1
#1 lore raw

Hi Kan,

Quoted 41 lines · On Thu, Jul 8, 2021 at 9:04 AM <[email protected]> wrote:

From: Kan Liang <[email protected]>

The Topdown Microarchitecture Analysis (TMA) Method is a structured analysis methodology to identify critical performance bottlenecks in out-of-order processors. The Topdown metrics L1 event was added as default events in the commit 42641d6f4d15 ("perf stat: Add Topdown metrics events as default events"). From the Sapphire Rapids server and later platforms, the same dedicated "metrics" register is extended to support both L1 and L2 events. Add both L1 and L2 Topdown metrics events as default events to enrich the default measuring information, if the new measurement register is available. On legacy systems there is no change to avoid extra multiplexing.

The topdown_level indicates the max metrics level for the top-down statistics. Set it to 2 to display all L1 and L2 Topdown metrics events.

With the patch:

 $ perf stat sleep 1

Performance counter stats for 'sleep 1':

              0.59 msec task-clock                #    0.001 CPUs utilized
                 1      context-switches          #    1.687 K/sec
                 0      cpu-migrations            #    0.000 /sec
                76      page-faults               #  128.198 K/sec
         1,405,318      cycles                    #    2.371 GHz
         1,471,136      instructions              #    1.05  insn per cycle
           310,132      branches                  #  523.136 M/sec
            10,435      branch-misses             #    3.36% of all branches
         8,431,908      slots                     #   14.223 G/sec
         1,554,116      topdown-retiring          #     18.4% retiring
         1,289,585      topdown-bad-spec          #     15.2% bad speculation
         2,810,636      topdown-fe-bound          #     33.2% frontend bound
         2,810,636      topdown-be-bound          #     33.2% backend bound
           231,464      topdown-heavy-ops         #      2.7% heavy operations       #     15.6% light operations
         1,223,453      topdown-br-mispredict     #     14.5% branch mispredict      #      0.8% machine clears
         1,884,779      topdown-fetch-lat         #     22.3% fetch latency          #     10.9% fetch bandwidth
         1,454,917      topdown-mem-bound         #     17.2% memory bound           #     16.0% Core bound
Quoted 7 lines
       1.001179699 seconds time elapsed

       0.000000000 seconds user
       0.001238000 seconds sys

Thanks,
Namhyung

Quoted 50 lines
---
 tools/perf/arch/x86/util/evlist.c | 6 +++++-
 tools/perf/builtin-stat.c         | 3 +++
 2 files changed, 8 insertions(+), 1 deletion(-)
tools/perf/arch/x86/util/evlist.c +5 -1
index 8c6732c..0b09510 100644
@@ -5,11 +5,15 @@
 #include "util/parse-events.h"
 
 #define TOPDOWN_L1_EVENTS      "{slots,topdown-retiring,topdown-bad-spec,topdown-fe-bound,topdown-be-bound}"
+#define TOPDOWN_L2_EVENTS      "{slots,topdown-retiring,topdown-bad-spec,topdown-fe-bound,topdown-be-bound,topdown-heavy-ops,topdown-br-mispredict,topdown-fetch-lat,topdown-mem-bound}"
 
 int arch_evlist__add_default_attrs(struct evlist *evlist)
 {
        if (!pmu_have_event("cpu", "slots"))
                return 0;
 
-       return parse_events(evlist, TOPDOWN_L1_EVENTS, NULL);
+       if (pmu_have_event("cpu", "topdown-heavy-ops"))
+               return parse_events(evlist, TOPDOWN_L2_EVENTS, NULL);
+       else
+               return parse_events(evlist, TOPDOWN_L1_EVENTS, NULL);
 }
tools/perf/builtin-stat.c +3 -1
index 5a830ae..a5e93d4 100644
@@ -154,6 +154,8 @@ static const char *topdown_metric_L2_attrs[] = {
        NULL,
 };
 
+#define TOPDOWN_MAX_LEVEL                      2
+
 static const char *smi_cost_attrs = {
        "{"
        "msr/aperf/,"
@@ -1924,6 +1926,7 @@ static int add_default_attributes(void)
                if (evlist__add_default_attrs(evsel_list, default_attrs1) < 0)
                        return -1;
 
+               stat_config.topdown_level = TOPDOWN_MAX_LEVEL;
                if (arch_evlist__add_default_attrs(evsel_list) < 0)
                        return -1;
        }
--

2.7.4

maintainer Applied
#2 lore raw

Em Thu, Jul 08, 2021 at 10:31:34PM -0700, Namhyung Kim escreveu:

Quoted 27 lines · On Thu, Jul 8, 2021 at 9:04 AM <[email protected]> wrote:
Quoted 25 lines

With the patch:

 $ perf stat sleep 1

Performance counter stats for 'sleep 1':

              0.59 msec task-clock                #    0.001 CPUs utilized
                 1      context-switches          #    1.687 K/sec
                 0      cpu-migrations            #    0.000 /sec
                76      page-faults               #  128.198 K/sec
         1,405,318      cycles                    #    2.371 GHz
         1,471,136      instructions              #    1.05  insn per cycle
           310,132      branches                  #  523.136 M/sec
            10,435      branch-misses             #    3.36% of all branches
         8,431,908      slots                     #   14.223 G/sec
         1,554,116      topdown-retiring          #     18.4% retiring
         1,289,585      topdown-bad-spec          #     15.2% bad speculation
         2,810,636      topdown-fe-bound          #     33.2% frontend bound
         2,810,636      topdown-be-bound          #     33.2% backend bound
           231,464      topdown-heavy-ops         #      2.7% heavy operations       #     15.6% light operations
         1,223,453      topdown-br-mispredict     #     14.5% branch mispredict      #      0.8% machine clears
         1,884,779      topdown-fetch-lat         #     22.3% fetch latency          #     10.9% fetch bandwidth
         1,454,917      topdown-mem-bound         #     17.2% memory bound           #     16.0% Core bound
       1.001179699 seconds time elapsed
Quoted 7 lines
Quoted 5 lines
       0.000000000 seconds user
       0.001238000 seconds sys

Thanks, applied.

- Arnaldo