Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1317145 > unrolled thread

'perf stat --repeat N' oddity/regression

Started byIngo Molnar <mingo@kernel.org>
First post2016-01-25 19:30 +0100
Last post2016-01-25 20:50 +0100
Articles 6 — 2 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  'perf stat --repeat N' oddity/regression Ingo Molnar <mingo@kernel.org> - 2016-01-25 19:30 +0100
    Re: 'perf stat --repeat N' oddity/regression Ingo Molnar <mingo@kernel.org> - 2016-01-25 19:40 +0100
      Re: 'perf stat --repeat N' oddity/regression Arnaldo Carvalho de Melo <acme@kernel.org> - 2016-01-25 20:50 +0100
        Re: 'perf stat --repeat N' oddity/regression Ingo Molnar <mingo@kernel.org> - 2016-01-26 09:10 +0100
          Re: 'perf stat --repeat N' oddity/regression Arnaldo Carvalho de Melo <acme@kernel.org> - 2016-01-26 15:20 +0100
      Re: 'perf stat --repeat N' oddity/regression Arnaldo Carvalho de Melo <acme@kernel.org> - 2016-01-25 20:50 +0100

#1317145 — 'perf stat --repeat N' oddity/regression

FromIngo Molnar <mingo@kernel.org>
Date2016-01-25 19:30 +0100
Subject'perf stat --repeat N' oddity/regression
Message-ID<qUU3N-830-33@gated-at.bofh.it>
So I just noticed this 'perf stat --repeat' oddity:

triton:~/tip> perf stat --repeat 3 -a -e instructions taskset 1 perf bench sched pipe -l 1000000
# Running 'sched/pipe' benchmark:
# Executed 1000000 pipe operations between two processes

     Total time: 2.068 [sec]

       2.068208 usecs/op
         483510 ops/sec
# Running 'sched/pipe' benchmark:
# Executed 1000000 pipe operations between two processes

     Total time: 2.097 [sec]

       2.097126 usecs/op
         476843 ops/sec
# Running 'sched/pipe' benchmark:
# Executed 1000000 pipe operations between two processes

     Total time: 2.091 [sec]

       2.091716 usecs/op
         478076 ops/sec

 Performance counter stats for 'system wide' (3 runs):

    12,690,156,407      instructions                                                

       2.087612423 seconds time elapsed                                          ( +-  0.43% )

the stddev printout of 'instructions' is gone! This defeats the main purpose of 
--repeat.

The 'elapsed' time +- stddev/noise display (which is still present) is only part 
of the story.

I'm pretty sure we had the printout for all the measured fields a couple of months 
ago - does any of you know what happened to it?

Thanks,

	Ingo

[toc] | [next] | [standalone]

#1317169

FromIngo Molnar <mingo@kernel.org>
Date2016-01-25 19:40 +0100
Message-ID<qUUdu-88K-59@gated-at.bofh.it>
In reply to#1317145
So it appears it broke somewhere between v4.0 and v4.1, as the v4.0 install 
displays:

    16,244,802,268      instructions               ( +- 23.01% )

       2.108676769 seconds time elapsed                                          ( +-  0.86% )


[a few minutes later]

Bisected it down to:

 106a94a0f8c207ef4113ce7e32f34a00b3b174e7 is the first bad commit
 commit 106a94a0f8c207ef4113ce7e32f34a00b3b174e7
 Author: Jiri Olsa <jolsa@kernel.org>
 Date:   Fri Jun 26 11:29:19 2015 +0200

     perf stat: Introduce read_counters function

So this look like to be a regression.

Thanks,

	Ingo

[toc] | [prev] | [next] | [standalone]

#1317247

FromArnaldo Carvalho de Melo <acme@kernel.org>
Date2016-01-25 20:50 +0100
Message-ID<qUVjd-pt-15@gated-at.bofh.it>
In reply to#1317169
Em Mon, Jan 25, 2016 at 04:43:33PM -0300, Arnaldo Carvalho de Melo escreveu:
> Em Mon, Jan 25, 2016 at 07:35:11PM +0100, Ingo Molnar escreveu:
> > 
> > So it appears it broke somewhere between v4.0 and v4.1, as the v4.0 install 
> > displays:
> > 
> >     16,244,802,268      instructions               ( +- 23.01% )
> > 
> >        2.108676769 seconds time elapsed                                          ( +-  0.86% )
> > 
> > 
> > [a few minutes later]
> > 
> > Bisected it down to:
> 
> Mel Gorman reported this and Jiri provided a patch that Mel tested and
> verified that it fixed, checking where it is sitting now...

I already sent it your way, its:

    198 N C 01/21 Arnaldo Carvalh (1.4K) ├─>[PATCH 06/16] perf stat: Do not clean event's private stats

Please pull my perf-core-for-mingo tag and you should get it, its an one
liner.

- Arnaldo
 
> - Arnaldo
>  
> >  106a94a0f8c207ef4113ce7e32f34a00b3b174e7 is the first bad commit
> >  commit 106a94a0f8c207ef4113ce7e32f34a00b3b174e7
> >  Author: Jiri Olsa <jolsa@kernel.org>
> >  Date:   Fri Jun 26 11:29:19 2015 +0200
> > 
> >      perf stat: Introduce read_counters function
> > 
> > So this look like to be a regression.
> > 
> > Thanks,
> > 
> > 	Ingo

[toc] | [prev] | [next] | [standalone]

#1317593

FromIngo Molnar <mingo@kernel.org>
Date2016-01-26 09:10 +0100
Message-ID<qV6Rk-1gU-13@gated-at.bofh.it>
In reply to#1317247
* Arnaldo Carvalho de Melo <acme@kernel.org> wrote:

> Em Mon, Jan 25, 2016 at 04:43:33PM -0300, Arnaldo Carvalho de Melo escreveu:
> > Em Mon, Jan 25, 2016 at 07:35:11PM +0100, Ingo Molnar escreveu:
> > > 
> > > So it appears it broke somewhere between v4.0 and v4.1, as the v4.0 install 
> > > displays:
> > > 
> > >     16,244,802,268      instructions               ( +- 23.01% )
> > > 
> > >        2.108676769 seconds time elapsed                                          ( +-  0.86% )
> > > 
> > > 
> > > [a few minutes later]
> > > 
> > > Bisected it down to:
> > 
> > Mel Gorman reported this and Jiri provided a patch that Mel tested and
> > verified that it fixed, checking where it is sitting now...
> 
> I already sent it your way, its:
> 
>     198 N C 01/21 Arnaldo Carvalh (1.4K) ├─>[PATCH 06/16] perf stat: Do not clean event's private stats
> 
> Please pull my perf-core-for-mingo tag and you should get it, its an one
> liner.

Hm, are all of those changes fixes? If not then mind splitting them into a 
perf/urgent portion as well?

Thanks!

	Ingo

[toc] | [prev] | [next] | [standalone]

#1317954

FromArnaldo Carvalho de Melo <acme@kernel.org>
Date2016-01-26 15:20 +0100
Message-ID<qVcDo-5i7-23@gated-at.bofh.it>
In reply to#1317593
Em Tue, Jan 26, 2016 at 09:08:22AM +0100, Ingo Molnar escreveu:
> 
> * Arnaldo Carvalho de Melo <acme@kernel.org> wrote:
> 
> > Em Mon, Jan 25, 2016 at 04:43:33PM -0300, Arnaldo Carvalho de Melo escreveu:
> > > Em Mon, Jan 25, 2016 at 07:35:11PM +0100, Ingo Molnar escreveu:
> > > > 
> > > > So it appears it broke somewhere between v4.0 and v4.1, as the v4.0 install 
> > > > displays:
> > > > 
> > > >     16,244,802,268      instructions               ( +- 23.01% )
> > > > 
> > > >        2.108676769 seconds time elapsed                                          ( +-  0.86% )
> > > > 
> > > > 
> > > > [a few minutes later]
> > > > 
> > > > Bisected it down to:
> > > 
> > > Mel Gorman reported this and Jiri provided a patch that Mel tested and
> > > verified that it fixed, checking where it is sitting now...
> > 
> > I already sent it your way, its:
> > 
> >     198 N C 01/21 Arnaldo Carvalh (1.4K) ├─>[PATCH 06/16] perf stat: Do not clean event's private stats
> > 
> > Please pull my perf-core-for-mingo tag and you should get it, its an one
> > liner.
> 
> Hm, are all of those changes fixes? If not then mind splitting them into a 
> perf/urgent portion as well?

Ok, doing that...

- Arnaldo

[toc] | [prev] | [next] | [standalone]

#1317255

FromArnaldo Carvalho de Melo <acme@kernel.org>
Date2016-01-25 20:50 +0100
Message-ID<qUVjd-pt-17@gated-at.bofh.it>
In reply to#1317169
Em Mon, Jan 25, 2016 at 07:35:11PM +0100, Ingo Molnar escreveu:
> 
> So it appears it broke somewhere between v4.0 and v4.1, as the v4.0 install 
> displays:
> 
>     16,244,802,268      instructions               ( +- 23.01% )
> 
>        2.108676769 seconds time elapsed                                          ( +-  0.86% )
> 
> 
> [a few minutes later]
> 
> Bisected it down to:

Mel Gorman reported this and Jiri provided a patch that Mel tested and
verified that it fixed, checking where it is sitting now...

- Arnaldo
 
>  106a94a0f8c207ef4113ce7e32f34a00b3b174e7 is the first bad commit
>  commit 106a94a0f8c207ef4113ce7e32f34a00b3b174e7
>  Author: Jiri Olsa <jolsa@kernel.org>
>  Date:   Fri Jun 26 11:29:19 2015 +0200
> 
>      perf stat: Introduce read_counters function
> 
> So this look like to be a regression.
> 
> Thanks,
> 
> 	Ingo

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web