Skip to contentSkip to conversation

perf stat: Improve runtime stat for interval mode

PATCH Superseded Jin Yao 3 participants 6 messages Apr 17, 2020 – Apr 20, 2020 perf events

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Jin Yao

author
#0 lore raw

PATCH 1/1 perf stat: Improve runtime stat for interval mode

For interval mode, the metric is printed after # if it exists. But it's not calculated by the counts generated in this interval. See following examples,

 root@kbl-ppc:~# perf stat -M CPI -I1000 --interval-count 2
 #           time             counts unit events
      1.000422803            764,809      inst_retired.any          #      2.9 CPI
      1.000422803          2,234,932      cycles
      2.001464585          1,960,061      inst_retired.any          #      1.6 CPI
      2.001464585          4,022,591      cycles

The second CPI should not be 1.6 (4,022,591/1,960,061 is 2.1)

 root@kbl-ppc:~# perf stat -e cycles,instructions -I1000 --interval-count 2
 #           time             counts unit events
      1.000429493          2,869,311      cycles
      1.000429493            816,875      instructions              #    0.28  insn per cycle
      2.001516426          9,260,973      cycles
      2.001516426          5,250,634      instructions              #    0.87  insn per cycle

The second 'insn per cycle' should not be 0.87 (5,250,634/9,260,973 is 0.57).

The current code uses a global variable rt_stat for tracking and updating the std dev of runtime stat. Unlike the counts, rt_stat is not reset for interval. While the counts are reset for interval.

perf_stat_process_counter()

{
	if (config->interval)
		init_stats(ps->res_stats);
}

So for interval, the rt_stat should be reset either.

This patch resets rt_stat before read_counters, so the runtime stat is only calculated by the counts generated in this interval.

With this patch,

 root@kbl-ppc:~# perf stat -M CPI -I1000 --interval-count 2
 #           time             counts unit events
      1.000420924          2,408,818      inst_retired.any          #      2.1 CPI
      1.000420924          5,010,111      cycles
      2.001448579          2,798,407      inst_retired.any          #      1.6 CPI
      2.001448579          4,599,861      cycles
 root@kbl-ppc:~# perf stat -e cycles,instructions -I1000 --interval-count 2
 #           time             counts unit events
      1.000428555          2,769,714      cycles
      1.000428555            774,462      instructions              #    0.28  insn per cycle
      2.001471562          3,595,904      cycles
      2.001471562          1,243,703      instructions              #    0.35  insn per cycle

Now the second 'insn per cycle' and CPI are calculated by the counts generated in this interval.

---
 tools/perf/Documentation/perf-stat.txt | 2 ++
 tools/perf/builtin-stat.c              | 1 +
 tools/perf/util/stat-shadow.c          | 5 +++++
 tools/perf/util/stat.h                 | 1 +
 4 files changed, 9 insertions(+)
tools/perf/Documentation/perf-stat.txt +2 -0
index 4d56586b2fb9..3fb5028aef08 100644
@@ -176,6 +176,8 @@ Print count deltas every N milliseconds (minimum: 1ms)
 The overhead percentage could be high in some cases, for instance with small, sub 100ms intervals.  Use with caution.
 	example: 'perf stat -I 1000 -e cycles -a sleep 5'
 
+If the metric exists, it is calculated by the counts generated in this interval and the metric is printed after #.
+
 --interval-count times::
 Print count deltas for fixed number of times.
 This option should be used together with "-I" option.
tools/perf/builtin-stat.c +1 -0
index ec053dc1e35c..d5c326ff46d0 100644
@@ -359,6 +359,7 @@ static void process_interval(void)
 	clock_gettime(CLOCK_MONOTONIC, &ts);
 	diff_timespec(&rs, &ts, &ref_time);
 
+	perf_stat__reset_rt_stat();
 	read_counters(&rs);
 
 	if (STAT_RECORD) {
tools/perf/util/stat-shadow.c +5 -0
index 1ad5c5be7e97..c8f5ca6c8749 100644
@@ -186,6 +186,11 @@ void perf_stat__reset_shadow_stats(void)
 	memset(&walltime_nsecs_stats, 0, sizeof(walltime_nsecs_stats));
 }
 
+void perf_stat__reset_rt_stat(void)
+{
+	reset_stat(&rt_stat);
+}
+
 void perf_stat__reset_shadow_per_stat(struct runtime_stat *st)
 {
 	reset_stat(st);
tools/perf/util/stat.h +1 -0
index b4fdfaa7f2c0..2975494166dd 100644
@@ -176,6 +176,7 @@ void runtime_stat__init(struct runtime_stat *st);
 void runtime_stat__exit(struct runtime_stat *st);
 void perf_stat__init_shadow_stats(void);
 void perf_stat__reset_shadow_stats(void);
+void perf_stat__reset_rt_stat(void);
 void perf_stat__reset_shadow_per_stat(struct runtime_stat *st);
 void perf_stat__update_shadow_stats(struct evsel *counter, u64 count,
 				    int cpu, struct runtime_stat *st);
Signature · 1 line
2.17.1
Acked-by on v1 1/1
#1 lore raw
Quoted 46 lines · Jin Yao

For interval mode, the metric is printed after # if it exists. But it's not calculated by the counts generated in this interval. See following examples,

 root@kbl-ppc:~# perf stat -M CPI -I1000 --interval-count 2
 #           time             counts unit events
      1.000422803            764,809      inst_retired.any          #      2.9 CPI
      1.000422803          2,234,932      cycles
      2.001464585          1,960,061      inst_retired.any          #      1.6 CPI
      2.001464585          4,022,591      cycles

The second CPI should not be 1.6 (4,022,591/1,960,061 is 2.1)

 root@kbl-ppc:~# perf stat -e cycles,instructions -I1000 --interval-count 2
 #           time             counts unit events
      1.000429493          2,869,311      cycles
      1.000429493            816,875      instructions              #    0.28  insn per cycle
      2.001516426          9,260,973      cycles
      2.001516426          5,250,634      instructions              #    0.87  insn per cycle

The second 'insn per cycle' should not be 0.87 (5,250,634/9,260,973 is 0.57).

The current code uses a global variable rt_stat for tracking and updating the std dev of runtime stat. Unlike the counts, rt_stat is not reset for interval. While the counts are reset for interval.

perf_stat_process_counter()

{
	if (config->interval)
		init_stats(ps->res_stats);
}

So for interval, the rt_stat should be reset either.

This patch resets rt_stat before read_counters, so the runtime stat is only calculated by the counts generated in this interval.

With this patch,

 root@kbl-ppc:~# perf stat -M CPI -I1000 --interval-count 2
 #           time             counts unit events
      1.000420924          2,408,818      inst_retired.any          #      2.1 CPI
      1.000420924          5,010,111      cycles
      2.001448579          2,798,407      inst_retired.any          #      1.6 CPI
      2.001448579          4,599,861      cycles
Quoted 7 lines
 root@kbl-ppc:~# perf stat -e cycles,instructions -I1000 --interval-count 2
 #           time             counts unit events
      1.000428555          2,769,714      cycles
      1.000428555            774,462      instructions              #    0.28  insn per cycle
      2.001471562          3,595,904      cycles
      2.001471562          1,243,703      instructions              #    0.35  insn per cycle

Hi Jin,
Thanks for the patch. Was working on similar issue raised by Joakim Zhang.
Here's link to my findings: https://lkml.org/lkml/2020/3/24/158

This patch solves the issue.

Quoted 64 lines

Now the second 'insn per cycle' and CPI are calculated by the counts generated in this interval.>

---
 tools/perf/Documentation/perf-stat.txt | 2 ++
 tools/perf/builtin-stat.c              | 1 +
 tools/perf/util/stat-shadow.c          | 5 +++++
 tools/perf/util/stat.h                 | 1 +
 4 files changed, 9 insertions(+)
tools/perf/Documentation/perf-stat.txt +2 -0
index 4d56586b2fb9..3fb5028aef08 100644
@@ -176,6 +176,8 @@ Print count deltas every N milliseconds (minimum: 1ms)
 The overhead percentage could be high in some cases, for instance with small, sub 100ms intervals.  Use with caution.
 	example: 'perf stat -I 1000 -e cycles -a sleep 5'
 
+If the metric exists, it is calculated by the counts generated in this interval and the metric is printed after #.
+
 --interval-count times::
 Print count deltas for fixed number of times.
 This option should be used together with "-I" option.
tools/perf/builtin-stat.c +1 -0
index ec053dc1e35c..d5c326ff46d0 100644
@@ -359,6 +359,7 @@ static void process_interval(void)
 	clock_gettime(CLOCK_MONOTONIC, &ts);
 	diff_timespec(&rs, &ts, &ref_time);
 
+	perf_stat__reset_rt_stat();
 	read_counters(&rs);
 
 	if (STAT_RECORD) {
tools/perf/util/stat-shadow.c +5 -0
index 1ad5c5be7e97..c8f5ca6c8749 100644
@@ -186,6 +186,11 @@ void perf_stat__reset_shadow_stats(void)
 	memset(&walltime_nsecs_stats, 0, sizeof(walltime_nsecs_stats));
 }
 
+void perf_stat__reset_rt_stat(void)
+{
+	reset_stat(&rt_stat);
+}
+
 void perf_stat__reset_shadow_per_stat(struct runtime_stat *st)
 {
 	reset_stat(st);
tools/perf/util/stat.h +1 -0
index b4fdfaa7f2c0..2975494166dd 100644
@@ -176,6 +176,7 @@ void runtime_stat__init(struct runtime_stat *st);
 void runtime_stat__exit(struct runtime_stat *st);
 void perf_stat__init_shadow_stats(void);
 void perf_stat__reset_shadow_stats(void);
+void perf_stat__reset_rt_stat(void);
 void perf_stat__reset_shadow_per_stat(struct runtime_stat *st);
 void perf_stat__update_shadow_stats(struct evsel *counter, u64 count,
 				    int cpu, struct runtime_stat *st);
Acked-by on v1 1/1
#2 lore raw
In reply to Jin Yao
Quoted 46 lines · Jin Yao

For interval mode, the metric is printed after # if it exists. But it's not calculated by the counts generated in this interval. See following examples,

 root@kbl-ppc:~# perf stat -M CPI -I1000 --interval-count 2
 #           time             counts unit events
      1.000422803            764,809      inst_retired.any          #      2.9 CPI
      1.000422803          2,234,932      cycles
      2.001464585          1,960,061      inst_retired.any          #      1.6 CPI
      2.001464585          4,022,591      cycles

The second CPI should not be 1.6 (4,022,591/1,960,061 is 2.1)

 root@kbl-ppc:~# perf stat -e cycles,instructions -I1000 --interval-count 2
 #           time             counts unit events
      1.000429493          2,869,311      cycles
      1.000429493            816,875      instructions              #    0.28  insn per cycle
      2.001516426          9,260,973      cycles
      2.001516426          5,250,634      instructions              #    0.87  insn per cycle

The second 'insn per cycle' should not be 0.87 (5,250,634/9,260,973 is 0.57).

The current code uses a global variable rt_stat for tracking and updating the std dev of runtime stat. Unlike the counts, rt_stat is not reset for interval. While the counts are reset for interval.

perf_stat_process_counter()

{
	if (config->interval)
		init_stats(ps->res_stats);
}

So for interval, the rt_stat should be reset either.

This patch resets rt_stat before read_counters, so the runtime stat is only calculated by the counts generated in this interval.

With this patch,

 root@kbl-ppc:~# perf stat -M CPI -I1000 --interval-count 2
 #           time             counts unit events
      1.000420924          2,408,818      inst_retired.any          #      2.1 CPI
      1.000420924          5,010,111      cycles
      2.001448579          2,798,407      inst_retired.any          #      1.6 CPI
      2.001448579          4,599,861      cycles
Quoted 7 lines
 root@kbl-ppc:~# perf stat -e cycles,instructions -I1000 --interval-count 2
 #           time             counts unit events
      1.000428555          2,769,714      cycles
      1.000428555            774,462      instructions              #    0.28  insn per cycle
      2.001471562          3,595,904      cycles
      2.001471562          1,243,703      instructions              #    0.35  insn per cycle

Hi Jin,
Thanks for the patch. Was working on similar issue raised by Joakim Zhang.
Here's link to my findings: https://lkml.org/lkml/2020/3/24/158

This patch solves the issue.

Quoted 65 lines

Now the second 'insn per cycle' and CPI are calculated by the counts generated in this interval.

---
 tools/perf/Documentation/perf-stat.txt | 2 ++
 tools/perf/builtin-stat.c              | 1 +
 tools/perf/util/stat-shadow.c          | 5 +++++
 tools/perf/util/stat.h                 | 1 +
 4 files changed, 9 insertions(+)
tools/perf/Documentation/perf-stat.txt +2 -0
index 4d56586b2fb9..3fb5028aef08 100644
@@ -176,6 +176,8 @@ Print count deltas every N milliseconds (minimum: 1ms)
 The overhead percentage could be high in some cases, for instance with small, sub 100ms intervals.  Use with caution.
 	example: 'perf stat -I 1000 -e cycles -a sleep 5'
 
+If the metric exists, it is calculated by the counts generated in this interval and the metric is printed after #.
+
 --interval-count times::
 Print count deltas for fixed number of times.
 This option should be used together with "-I" option.
tools/perf/builtin-stat.c +1 -0
index ec053dc1e35c..d5c326ff46d0 100644
@@ -359,6 +359,7 @@ static void process_interval(void)
 	clock_gettime(CLOCK_MONOTONIC, &ts);
 	diff_timespec(&rs, &ts, &ref_time);
 
+	perf_stat__reset_rt_stat();
 	read_counters(&rs);
 
 	if (STAT_RECORD) {
tools/perf/util/stat-shadow.c +5 -0
index 1ad5c5be7e97..c8f5ca6c8749 100644
@@ -186,6 +186,11 @@ void perf_stat__reset_shadow_stats(void)
 	memset(&walltime_nsecs_stats, 0, sizeof(walltime_nsecs_stats));
 }
 
+void perf_stat__reset_rt_stat(void)
+{
+	reset_stat(&rt_stat);
+}
+
 void perf_stat__reset_shadow_per_stat(struct runtime_stat *st)
 {
 	reset_stat(st);
tools/perf/util/stat.h +1 -0
index b4fdfaa7f2c0..2975494166dd 100644
@@ -176,6 +176,7 @@ void runtime_stat__init(struct runtime_stat *st);
 void runtime_stat__exit(struct runtime_stat *st);
 void perf_stat__init_shadow_stats(void);
 void perf_stat__reset_shadow_stats(void);
+void perf_stat__reset_rt_stat(void);
 void perf_stat__reset_shadow_per_stat(struct runtime_stat *st);
 void perf_stat__update_shadow_stats(struct evsel *counter, u64 count,
 				    int cpu, struct runtime_stat *st);
author
#3 lore raw
Quoted 57 lines · kajoljain
Quoted 46 lines · Jin Yao

For interval mode, the metric is printed after # if it exists. But it's not calculated by the counts generated in this interval. See following examples,

  root@kbl-ppc:~# perf stat -M CPI -I1000 --interval-count 2
  #           time             counts unit events
       1.000422803            764,809      inst_retired.any          #      2.9 CPI
       1.000422803          2,234,932      cycles
       2.001464585          1,960,061      inst_retired.any          #      1.6 CPI
       2.001464585          4,022,591      cycles

The second CPI should not be 1.6 (4,022,591/1,960,061 is 2.1)

  root@kbl-ppc:~# perf stat -e cycles,instructions -I1000 --interval-count 2
  #           time             counts unit events
       1.000429493          2,869,311      cycles
       1.000429493            816,875      instructions              #    0.28  insn per cycle
       2.001516426          9,260,973      cycles
       2.001516426          5,250,634      instructions              #    0.87  insn per cycle

The second 'insn per cycle' should not be 0.87 (5,250,634/9,260,973 is 0.57).

The current code uses a global variable rt_stat for tracking and updating the std dev of runtime stat. Unlike the counts, rt_stat is not reset for interval. While the counts are reset for interval.

perf_stat_process_counter()

{
	if (config->interval)
		init_stats(ps->res_stats);
}

So for interval, the rt_stat should be reset either.

This patch resets rt_stat before read_counters, so the runtime stat is only calculated by the counts generated in this interval.

With this patch,

  root@kbl-ppc:~# perf stat -M CPI -I1000 --interval-count 2
  #           time             counts unit events
       1.000420924          2,408,818      inst_retired.any          #      2.1 CPI
       1.000420924          5,010,111      cycles
       2.001448579          2,798,407      inst_retired.any          #      1.6 CPI
       2.001448579          4,599,861      cycles
Quoted 7 lines
  root@kbl-ppc:~# perf stat -e cycles,instructions -I1000 --interval-count 2
  #           time             counts unit events
       1.000428555          2,769,714      cycles
       1.000428555            774,462      instructions              #    0.28  insn per cycle
       2.001471562          3,595,904      cycles
       2.001471562          1,243,703      instructions              #    0.35  insn per cycle
Quoted 8 lines

Hi Jin,
Thanks for the patch. Was working on similar issue raised by Joakim Zhang.
Here's link to my findings: https://lkml.org/lkml/2020/3/24/158

This patch solves the issue.

Thanks for testing this patch. :)

Thanks
Jin Yao

Quoted 65 lines
Quoted 65 lines

Now the second 'insn per cycle' and CPI are calculated by the counts generated in this interval.

---
  tools/perf/Documentation/perf-stat.txt | 2 ++
  tools/perf/builtin-stat.c              | 1 +
  tools/perf/util/stat-shadow.c          | 5 +++++
  tools/perf/util/stat.h                 | 1 +
  4 files changed, 9 insertions(+)
tools/perf/Documentation/perf-stat.txt +2 -0
index 4d56586b2fb9..3fb5028aef08 100644
@@ -176,6 +176,8 @@ Print count deltas every N milliseconds (minimum: 1ms)
  The overhead percentage could be high in some cases, for instance with small, sub 100ms intervals.  Use with caution.
  	example: 'perf stat -I 1000 -e cycles -a sleep 5'
 
+If the metric exists, it is calculated by the counts generated in this interval and the metric is printed after #.
+
  --interval-count times::
  Print count deltas for fixed number of times.
  This option should be used together with "-I" option.
tools/perf/builtin-stat.c +1 -0
index ec053dc1e35c..d5c326ff46d0 100644
@@ -359,6 +359,7 @@ static void process_interval(void)
  	clock_gettime(CLOCK_MONOTONIC, &ts);
  	diff_timespec(&rs, &ts, &ref_time);
 
+	perf_stat__reset_rt_stat();
  	read_counters(&rs);
 
  	if (STAT_RECORD) {
tools/perf/util/stat-shadow.c +5 -0
index 1ad5c5be7e97..c8f5ca6c8749 100644
@@ -186,6 +186,11 @@ void perf_stat__reset_shadow_stats(void)
  	memset(&walltime_nsecs_stats, 0, sizeof(walltime_nsecs_stats));
  }
 
+void perf_stat__reset_rt_stat(void)
+{
+	reset_stat(&rt_stat);
+}
+
  void perf_stat__reset_shadow_per_stat(struct runtime_stat *st)
  {
  	reset_stat(st);
tools/perf/util/stat.h +1 -0
index b4fdfaa7f2c0..2975494166dd 100644
@@ -176,6 +176,7 @@ void runtime_stat__init(struct runtime_stat *st);
  void runtime_stat__exit(struct runtime_stat *st);
  void perf_stat__init_shadow_stats(void);
  void perf_stat__reset_shadow_stats(void);
+void perf_stat__reset_rt_stat(void);
  void perf_stat__reset_shadow_per_stat(struct runtime_stat *st);
  void perf_stat__update_shadow_stats(struct evsel *counter, u64 count,
  				    int cpu, struct runtime_stat *st);
#4 lore raw
In reply to Jin Yao

On Fri, Apr 17, 2020 at 08:51:54AM +0800, Jin Yao wrote:

SNIP

tools/perf/builtin-stat.c static void process_interval(void) · line 362

Quoted 9 lines
tools/perf/builtin-stat.c +1 -0
index ec053dc1e35c..d5c326ff46d0 100644
@@ -359,6 +359,7 @@ static void process_interval(void)
 	clock_gettime(CLOCK_MONOTONIC, &ts);
 	diff_timespec(&rs, &ts, &ref_time);
 
+	perf_stat__reset_rt_stat();

could you use just existing:

perf_stat__reset_shadow_per_stat(&rt_stat)

or perhaps event perf_stat__reset_shadow_stats, given that we init walltime_nsecs_stats just few lines below

jirka

Quoted 3 lines
tools/perf/builtin-stat.c +0 -0
 	read_counters(&rs);
 
 	if (STAT_RECORD) {

SNIP

author
#5 lore raw

Hi Jiri,

Quoted 16 lines · Jiri Olsa

On Fri, Apr 17, 2020 at 08:51:54AM +0800, Jin Yao wrote:

SNIP

tools/perf/builtin-stat.c static void process_interval(void) · line 362

Quoted 9 lines
tools/perf/builtin-stat.c +1 -0
index ec053dc1e35c..d5c326ff46d0 100644
@@ -359,6 +359,7 @@ static void process_interval(void)
  	clock_gettime(CLOCK_MONOTONIC, &ts);
  	diff_timespec(&rs, &ts, &ref_time);
 
+	perf_stat__reset_rt_stat();

could you use just existing:

Quoted 7 lines

perf_stat__reset_shadow_per_stat(&rt_stat)

or perhaps event perf_stat__reset_shadow_stats, given that we init walltime_nsecs_stats just few lines below

jirka

Yes, I can use the existing function to reset rt_stat, I don't need to define a new one.

Posted v2 yet.

Thanks
Jin Yao

Quoted 6 lines
Quoted 3 lines
tools/perf/builtin-stat.c +0 -0
  	read_counters(&rs);
 
  	if (STAT_RECORD) {

SNIP