Re: [PATCH 0/4] perf tools: New comm infrastructure

[Arnaldo Carvalho de Melo <acme@redhat.com>]

Em Tue, Sep 17, 2013 at 02:54:22PM +0900, Namhyung Kim escreveu:
> Hi Ingo,
> 
> On Sat, 14 Sep 2013 08:11:49 +0200, Ingo Molnar wrote:
> > * Frederic Weisbecker <fweisbec@gmail.com> wrote:
> >> My patches and Namhyung's should improve the comm situation a lot but we 
> >> can't do much miracle. The only way would be perhaps to be able to limit 
> >> the deepness of the callchain branches.
> >> 
> >> Now may be we can find other big contention point in perf. It's possible 
> >> we also have some endless loop somewhere.
> >
> > Well, it was the 100,000+ step linear list walk that was causing 90% of 
> > the slowness here. Namhyung's patch should dramatically improve that. I 
> > guess time for someone to post a combined tree so that it can be tested 
> > all together?
> 
> I pushed combined tree to 'perf/callchain-v2' branch in my tree
> 
>   git://git.kernel.org/pub/scm/linux/kernel/git/namhyung/linux-perf.git

But you did it for some old tree, as I tried running it now and since my
big perf.data file has PERF_RECORD_MMAP2 records, I can't try it with
your branch, working on cherry picking those changesets, since a plain
rebase didn't work.

- Arnaldo
 
> 
> Please note that I also pushed other versions (v[1-3]).  The v1 is my
> previous rbtree conversion patch, v2 adds Frederic's new comm
> infrastructure series on top and v3 adds my revised patch to refer
> current comm [1] on top of v2.
> 
> I did my own test again among them.  Test data is 400MB perf.data file
> created by parallel kernel build.
> 
>   $ ls -lh perf.data.big
>   -rw-------. 1 namhyung namhyung 400M Sep  9 10:21 perf.data.big
> 
> For more precise result, I changed cpufreq governor to 'performance'
> 
>   # echo performance > /sys/devices/system/cpu/cpu3/cpufreq/scaling_governor
> 
> and run perf report on the cpu.
> 
>   $ taskset -c 3 time -p perf --no-pager report --stdio -i perf.data.big > /dev/null
> 
> I ran it multiple times for each case and the results did not vary much.
> 
>         baseline	    v1            v2              v3
>   ----------------------------------------------------------
>   real    380.17	 12.63 	       10.02		9.03
>   user    378.86	 11.95 	        9.66		8.69
>   sys       0.70	  0.65 	        0.33		0.34
> 
> 
> I also tried to cache latest result and reuse it when adding a callchain
> (in callchain_append() function) but it only hits ~5% and did not help
> the performance.
> 
> Thanks,
> Namhyung
> 
> 
> [1] https://lkml.org/lkml/2013/9/16/565