Re: [PATCH 1/8] perf callchain: Convert children list to rbtree

[Namhyung Kim <namhyung@kernel.org>]

On Wed, 2 Oct 2013 12:18:28 +0200, Frederic Weisbecker wrote:
> On Thu, Sep 26, 2013 at 05:58:03PM +0900, Namhyung Kim wrote:
>> From: Namhyung Kim <namhyung.kim@lge.com>
>> 
>> Current collapse stage has a scalability problem which can be
>> reproduced easily with parallel kernel build.  This is because it
>> needs to traverse every children of callchain linearly during the
>> collapse/merge stage.  Convert it to rbtree reduced the overhead
>> significantly.
>> 
>> On my 400MB perf.data file which recorded with make -j32 kernel build:
>> 
>>   $ time perf --no-pager report --stdio > /dev/null
>> 
>> before:
>>   real	6m22.073s
>>   user	6m18.683s
>>   sys	0m0.706s
>> 
>> after:
>>   real	0m20.780s
>>   user	0m19.962s
>>   sys	0m0.689s
>> 
>> During the perf report the overhead on append_chain_children went down
>> from 96.69% to 18.16%:
>> 
>>   -  18.16%  perf  perf                [.] append_chain_children
>>      - append_chain_children
>>         - 77.48% append_chain_children
>>            + 69.79% merge_chain_branch
>>            - 22.96% append_chain_children
>>               + 67.44% merge_chain_branch
>>               + 30.15% append_chain_children
>>               + 2.41% callchain_append
>>            + 7.25% callchain_append
>>         + 12.26% callchain_append
>>         + 10.22% merge_chain_branch
>>   +  11.58%  perf  perf                [.] dso__find_symbol
>>   +   8.02%  perf  perf                [.] sort__comm_cmp
>>   +   5.48%  perf  libc-2.17.so        [.] malloc_consolidate
>> 
>> Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
>> Cc: Jiri Olsa <jolsa@redhat.com>
>> Cc: Frederic Weisbecker <fweisbec@gmail.com>
>> Link: http://lkml.kernel.org/n/tip-d9tcfow6stbrp4btvgs51y67@git.kernel.org
>> Signed-off-by: Namhyung Kim <namhyung@kernel.org>
>
> Have you tested this patchset when collapsing is not used?
> There are fair chances that this patchset does not only improve collapsing
> but also callchain insertion in general. So it's probably a win in any case. But
> still it would be nice to make sure that it's the case because we are getting
> rid of collapsing anyway.
>
> The test that could tell us about that is to run "perf report -s sym" and compare the
> time it takes to complete before and after this patch, because "-s sym" shouldn't
> involve collapses.
>
> Sorting by anything that is not comm should do the trick in fact.

Yes, I have similar result when collapsing is not used.  Actually when I
ran "perf report -s sym", the performance improves higher since it'd
insert more callchains in a hist entry.

Thanks,
Namhyung