Message-ID: <BANLkTikSMR7AyJi-LEr8iA2x9TxobJr9gA@mail.gmail.com>
Date:	Tue, 24 May 2011 16:05:53 -0600
From:	Jim Cromie <jim.cromie@...il.com>
To:	Ingo Molnar <mingo@...e.hu>
Cc:	linux-kernel@...r.kernel.org, a.p.zijlstra@...llo.nl,
	paulus@...ba.org, acme@...stprotocols.net
Subject: Re: [PATCH 2/2] dont commify big numbers by default, let -B do it

On Sat, May 21, 2011 at 4:33 AM, Ingo Molnar <mingo@...e.hu> wrote:
>
> * Jim Cromie <jim.cromie@...il.com> wrote:
>
>> Currently, big numbers get commas by default, which complicates parsing
>> by automation. [...]
>
> Well, automation can turn it off via --no-big-num, right?

ah.  my man-page is a bit behind
(but is in usage, and in Doc/perf-stat.txt)

> Also, i think for automation we'd also like to have a 'simple output' mode,
> would you like to add that?

OK, heres 1st cut at it, adding option --simple, for review / feedback

Its based upon the csv-output code mostly, with some vestiges of verbose..

I altered csv-output to have event-name 1st
this seems what folks would expect normally,
I surmised that existing was a short-cut..
Then added "noise"

part of reason for sending early is this column swap,
it might annoy existing users.

[jimc@...ucho perf]$ ./perf stat -r3 -x' ' true
task-clock-msecs1.009533  0.142%
context-switches 0 -nan%
CPU-migrations 0 100.000%
page-faults 85 0.000%
cycles 789434 0.107%
instructions 488768 0.546%
branches 102796 0.505%
branch-misses 8472 1.688%
cache-references 0 -nan%  (scaled from 0.00%)
cache-misses 0 -nan%  (scaled from 0.00%)

[jimc@...ucho perf]$ ./perf stat -r3 -x' ------- ' true
task-clock-msecs1.036450 -------  ------- 3.519%
context-switches ------- 0 ------- -nan%
CPU-migrations ------- 0 ------- -nan%
page-faults ------- 85 ------- 0.000%
cycles ------- 812151 ------- 3.463%
instructions ------- 499263 ------- 0.368%
branches ------- 104240 ------- 0.224%
branch-misses ------- 8125 ------- 4.627%
cache-references ------- 0 ------- -nan%  (scaled from 0.00%)
cache-misses ------- 0 ------- -nan%  (scaled from 0.00%)

I also need to do something better with -no-aggr format,
doesnt do right with extra column

[jimc@...ucho perf]$ sudo ./perf stat -r3 -x' ------- ' -A -a true
CPU0 ------- task-clock-msecs2.099862 -------
CPU1 ------- task-clock-msecs2.084522 -------
CPU2 ------- task-clock-msecs2.072948 -------
CPU3 ------- task-clock-msecs2.071611 -------
CPU0 ------- context-switches ------- 2
CPU1 ------- context-switches ------- 0
CPU2 ------- context-switches ------- 3
CPU3 ------- context-switches ------- 6
CPU0 ------- CPU-migrations ------- 2
...




> Something a bit like what you can see in 'perf stat -v true':
> Without the human output later on, and with elapsed time added as well.

for my part, Id like the moral equivalent of time(s)? output too,
though I suspect thats a separate patch..

[jimc@...ucho perf]$ time ./perf stat -x'  ' -- sh -c 'sleep 3'
task-clock-msecs6.830999
context-switches  2
CPU-migrations  1
page-faults  477
cycles  5404163  (scaled from 72.07%)
instructions  3302699
branches  804750
branch-misses  51870
cache-references  1584532  (scaled from 45.93%)
cache-misses  36616  (scaled from 31.36%)

real	0m3.019s
user	0m0.005s
sys	0m0.012s

are these timings already taken by perf-stat ?
is it a simple matter of addition and printing ?
If not, whats involved ?

Also, task-clock-msec doesnt quite match up with times' user number
Whats going on here ?


> Thanks,
>
>        Ingo
>

thank you
~jimc

PS: attaching now (in gmail web iface),
will try to thread up to this, RSN, with git send-email

View attachment "0001-add-simple-output-mode-to-perf-stat-based-upon-csv-o.patch" of type "text/x-patch" (5447 bytes)

Powered by blists - more mailing lists