LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Arnaldo Carvalho de Melo <acme@kernel.org>
To: Ingo Molnar <mingo@kernel.org>
Cc: Clark Williams <williams@redhat.com>,
	linux-kernel@vger.kernel.org, linux-perf-users@vger.kernel.org,
	Jiri Olsa <jolsa@kernel.org>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Andi Kleen <ak@linux.intel.com>, David Ahern <dsahern@gmail.com>,
	Namhyung Kim <namhyung@kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Arnaldo Carvalho de Melo <acme@redhat.com>
Subject: [PATCH 05/52] perf stat: Display time in precision based on std deviation
Date: Wed,  2 May 2018 12:18:55 -0300	[thread overview]
Message-ID: <20180502151942.20542-6-acme@kernel.org> (raw)
In-Reply-To: <20180502151942.20542-1-acme@kernel.org>

From: Jiri Olsa <jolsa@kernel.org>

Ingo suggested to display elapsed time for multirun workload (perf stat
-e) with precision based on the precision of the standard deviation.

In his own words:

  > This output is a slightly bit misleading:

  >  Performance counter stats for 'make -j128' (10 runs):
  >      27.988995256 seconds time elapsed                  ( +-  0.39% )

  > The 9 significant digits in the result, while only 1 is valid, suggests accuracy
  > where none exists.

  > It would be better if 'perf stat' would display elapsed time with a precision
  > adjusted to stddev, it should display at most 2 more significant digits than
  > the stddev inaccuracy.

  > I.e. in the above case 0.39% is 0.109, so we only have accuracy for 1 digit, and
  > so we should only display 3:

  >        27.988 seconds time elapsed                       ( +-  0.39% )

Plus a suggestion about the output, which is small enough and connected
with the above change that I merged both changes together.

  > Small output style nit - I think it would be nice if with --repeat the stddev was
  > also displayed in absolute values, besides percentage:
  >
  >       27.988 +- 0.109 seconds time elapsed   ( +- 0.39% )

The output is now:

   Performance counter stats for './perf bench sched pipe' (5 runs):
   SNIP
           13.3667 +- 0.0256 seconds time elapsed  ( +-  0.19% )

Suggested-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20180423090823.32309-7-jolsa@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-stat.c | 30 ++++++++++++++++++++++++------
 1 file changed, 24 insertions(+), 6 deletions(-)

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index f17dc601b0f3..3a94647ef6bc 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -1764,19 +1764,37 @@ static void print_header(int argc, const char **argv)
 	}
 }
 
+static int get_precision(double num)
+{
+	if (num > 1)
+		return 0;
+
+	return lround(ceil(-log10(num)));
+}
+
 static void print_footer(void)
 {
+	double avg = avg_stats(&walltime_nsecs_stats) / NSEC_PER_SEC;
 	FILE *output = stat_config.output;
 	int n;
 
 	if (!null_run)
 		fprintf(output, "\n");
-	fprintf(output, " %17.9f seconds time elapsed",
-			avg_stats(&walltime_nsecs_stats) / NSEC_PER_SEC);
-	if (run_count > 1) {
-		fprintf(output, "                                        ");
-		print_noise_pct(stddev_stats(&walltime_nsecs_stats),
-				avg_stats(&walltime_nsecs_stats));
+
+	if (run_count == 1) {
+		fprintf(output, " %17.9f seconds time elapsed", avg);
+	} else {
+		double sd = stddev_stats(&walltime_nsecs_stats) / NSEC_PER_SEC;
+		/*
+		 * Display at most 2 more significant
+		 * digits than the stddev inaccuracy.
+		 */
+		int precision = get_precision(sd) + 2;
+
+		fprintf(output, " %17.*f +- %.*f seconds time elapsed",
+			precision, avg, precision, sd);
+
+		print_noise_pct(sd, avg);
 	}
 	fprintf(output, "\n\n");
 
-- 
2.14.3

  parent reply	other threads:[~2018-05-02 15:20 UTC|newest]

Thread overview: 53+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-02 15:18 [GIT PULL 00/52] perf/core improvements and fixes Arnaldo Carvalho de Melo
2018-05-02 15:18 ` [PATCH 01/52] perf buildid-cache: Support --list option Arnaldo Carvalho de Melo
2018-05-02 15:18 ` [PATCH 02/52] perf buildid-cache: Support --purge-all option Arnaldo Carvalho de Melo
2018-05-02 15:18 ` [PATCH 03/52] perf check-headers.sh: Simplify arguments passing Arnaldo Carvalho de Melo
2018-05-02 15:18 ` [PATCH 04/52] perf check-headers.sh: Add support to check 2 independent files Arnaldo Carvalho de Melo
2018-05-02 15:18 ` Arnaldo Carvalho de Melo [this message]
2018-05-02 15:18 ` [PATCH 06/52] perf stat: Add --table option to display time of each run Arnaldo Carvalho de Melo
2018-05-02 15:18 ` [PATCH 07/52] perf stat: Display length strings of each run for --table option Arnaldo Carvalho de Melo
2018-05-02 15:18 ` [PATCH 08/52] perf top: Use __map__is_kernel() Arnaldo Carvalho de Melo
2018-05-02 15:18 ` [PATCH 09/52] perf symbols: Use __map__is_kernel() instead of ad-hoc equivalent code Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 10/52] perf dso: Add dso__has_symbols() method Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 11/52] perf map: Introduce map__has_symbols() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 12/52] perf thread: Introduce thread__find_map() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 13/52] perf tests: Let 'perf test list' display subtests Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 14/52] perf thread: Introduce thread__find_symbol() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 15/52] perf script: Use thread__find_symbol() instead of ad-hoc equivalent Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 16/52] perf thread: Make thread__find_map() return the map Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 17/52] perf thread: Make thread__find_symbol() return the symbol searched Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 18/52] perf map: Shorten map_groups__find_by_name() signature Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 19/52] perf Documentation: Support for asciidoctor Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 20/52] perf machine: Introduce machine__kernel_maps() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 21/52] perf machine: Shorten machine__load_kallsyms() signature Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 22/52] perf machine: Remove needless map_type from machine__load_vmlinux_path() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 23/52] perf tests vmlinux-kallsyms: Use machine__find_kernel_function(_by_name) Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 24/52] perf tests vmlinux-kallsyms: Use map__for_each_symbol() instead of open coded equivalent Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 25/52] perf map: Shorten map_groups__find() signature Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 26/52] perf ui stdio: Use map_groups__fprintf() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 27/52] perf symbols: Shorten dso__(first|last)_symbol() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 28/52] tools lib symbols: Introduce kallsyms__is_function() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 29/52] perf tools: Use kallsyms__is_function() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 30/52] perf symbols: Unexport symbol_type__is_a() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 31/52] perf map: Remove enum_type arg to map_groups__first() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 32/52] perf symbols: Remove map_type arg from dso__find_symbol() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 33/52] perf thread: Remove addr_type arg from thread__find_cpumode_addr_location() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 34/52] perf machine: Use machine__find_kernel_function() instead of open coded version Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 35/52] perf thread: Ditch __thread__find_symbol() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 36/52] perf thread: Make thread__find_map() search all maps Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 37/52] perf map: Remove map_type arg from map_groups__find() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 38/52] perf symbols: Store the ELF symbol type in the symbol struct Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 39/52] perf machine: Set PROT_EXEC for executable PERF_RECORD_MMAP records Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 40/52] perf sort: Use mmap->prot on "dcacheline" formatting Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 41/52] perf symbols: No need to special case MAP__FUNCTION in fixup Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 42/52] perf symbols: Use symbol type instead of map->type Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 43/52] perf map: Use map->prot in place of type==MAP__FUNCTION Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 44/52] perf symbols: " Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 45/52] perf symbols: Unify symbol maps Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 46/52] perf symbols: Remove needless goto Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 47/52] perf symbols: Split kernel symbol processing from dso__load_sym() Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 48/52] perf symbols: Remove unused dso__load_all_kallsyms() 'map' parameter Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 49/52] perf symbols: kallsyms__delta() needs the kmap, not the map Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 50/52] perf symbols: Move split_kallsyms to struct map_groups Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 51/52] perf tools: Fix spelling mistake: "builid" -> "buildid" Arnaldo Carvalho de Melo
2018-05-02 15:19 ` [PATCH 52/52] perf machine: Ditch find_kernel_function variants Arnaldo Carvalho de Melo

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180502151942.20542-6-acme@kernel.org \
    --to=acme@kernel.org \
    --cc=acme@redhat.com \
    --cc=ak@linux.intel.com \
    --cc=alexander.shishkin@linux.intel.com \
    --cc=dsahern@gmail.com \
    --cc=jolsa@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-perf-users@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=namhyung@kernel.org \
    --cc=peterz@infradead.org \
    --cc=williams@redhat.com \
    --subject='Re: [PATCH 05/52] perf stat: Display time in precision based on std deviation' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).