[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20090810140507.GA13540@elte.hu>
Date: Mon, 10 Aug 2009 16:05:07 +0200
From: Ingo Molnar <mingo@...e.hu>
To: mingo@...hat.com, hpa@...or.com, Brice.Goglin@...ia.fr,
linux-kernel@...r.kernel.org, peterz@...radead.org,
tglx@...utronix.de
Cc: linux-tip-commits@...r.kernel.org
Subject: Re: [tip:perfcounters/core] perf report: Add raw displaying of
per-thread counters
* tip-bot for Brice Goglin <Brice.Goglin@...ia.fr> wrote:
> Commit-ID: 9f8666971185b86615a074bcac67c90fdf8af8bc
> Gitweb: http://git.kernel.org/tip/9f8666971185b86615a074bcac67c90fdf8af8bc
> Author: Brice Goglin <Brice.Goglin@...ia.fr>
> AuthorDate: Mon, 10 Aug 2009 15:26:32 +0200
> Committer: Ingo Molnar <mingo@...e.hu>
> CommitDate: Mon, 10 Aug 2009 15:48:17 +0200
>
> perf report: Add raw displaying of per-thread counters
>
> If --pretty=raw is given to perf report -T, it now displays one
> line per-thread per-counter with the raw event id added.
>
> We get:
> # PID TID Name Raw Count
> 18608 18609 cache-misses 28e 416744
> 18608 18609 cache-references 28f 6456792
> 18608 18608 cache-misses 28e 448219
> 18608 18608 cache-references 28f 7270244
btw., this is the full output we do:
aldebaran:~> perf record -f -n -s ./hackbench 1
aldebaran:~> perf report -T --pretty=raw
# Samples: 0
#
# Overhead Command Shared Object Symbol
# ........ ......... ............. ......
#
#
# (For a higher level overview, try: perf report --sort comm,dso)
#
# PID TID Name Raw Count
17937 17937 cycles 3b 11590867
17935 17935 cycles 3b 11461700
17942 17942 cycles 3b 10912077
17927 17927 cycles 3b 10904967
17924 17924 cycles 3b 13313815
17934 17934 cycles 3b 11639158
17939 17939 cycles 3b 10803818
17931 17931 cycles 3b 11802979
17925 17925 cycles 3b 11303473
17926 17926 cycles 3b 12243882
17936 17936 cycles 3b 9997099
17938 17938 cycles 3b 11255911
17932 17932 cycles 3b 11957781
17929 17929 cycles 3b 12936176
17933 17933 cycles 3b 12685907
17930 17930 cycles 3b 12455278
17928 17928 cycles 3b 13187215
17940 17940 cycles 3b 15426327
17923 17923 cycles 3b 15754461
17903 17903 cycles 3b 11156016
17905 17905 cycles 3b 9747782
17904 17904 cycles 3b 8743904
17907 17907 cycles 3b 10827088
17906 17906 cycles 3b 10109984
17908 17908 cycles 3b 11092446
17909 17909 cycles 3b 8167235
17910 17910 cycles 3b 12132769
17911 17911 cycles 3b 9469716
17912 17912 cycles 3b 9541507
17914 17914 cycles 3b 8612516
17913 17913 cycles 3b 9805519
17915 17915 cycles 3b 10021933
17916 17916 cycles 3b 10763787
17917 17917 cycles 3b 10519547
17918 17918 cycles 3b 9749113
17919 17919 cycles 3b 10582943
17921 17921 cycles 3b 8717079
17920 17920 cycles 3b 9113000
17922 17922 cycles 3b 12102473
17941 17941 cycles 3b 15668897
17902 17902 cycles 3b 9148959
I think we should supress the header portion, as it's highly
confusing in this case when we dont sample RIPs but only counts,
right?
Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists