Comparing output of two perf cpu profiling reports

Comparing output of two perf cpu profiling reports

[Mark Davis]

Hi, what is the recommend way to compare two different perf reports
from CPU event profiling (via sampling with perf record)? I'm
comparing two macro benchmarks, with a couple small implementation
details between the two. Right now I'm manually comparing the output
of perf report to determine the salient differences. However, I'm
wondering if there's a tool or otherwise best practice to essentially
"diff" the two outputs. I could just use diff on the --stdio version
of the reports, but sometimes lines get mangled and I loose
context/hierarchy.

Any tips?

Mark

Re: Comparing output of two perf cpu profiling reports

[Alex Bagehot]

Hi Mark,

Differential flame graphs is one way:
http://www.brendangregg.com/blog/2014-11-09/differential-flame-graphs.html

Thanks,
Alex

On Thu, Apr 28, 2016 at 2:27 PM, Mark Davis <markdavisinboston@gmail.com> wrote:
> Hi, what is the recommend way to compare two different perf reports
> from CPU event profiling (via sampling with perf record)? I'm
> comparing two macro benchmarks, with a couple small implementation
> details between the two. Right now I'm manually comparing the output
> of perf report to determine the salient differences. However, I'm
> wondering if there's a tool or otherwise best practice to essentially
> "diff" the two outputs. I could just use diff on the --stdio version
> of the reports, but sometimes lines get mangled and I loose
> context/hierarchy.
>
> Any tips?
>
> Mark
> --
> To unsubscribe from this list: send the line "unsubscribe linux-perf-users" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Comparing output of two perf cpu profiling reports

[Arnaldo Carvalho de Melo]

Em Thu, Apr 28, 2016 at 09:27:07AM -0400, Mark Davis escreveu:
> Hi, what is the recommend way to compare two different perf reports
> from CPU event profiling (via sampling with perf record)?

> I'm comparing two macro benchmarks, with a couple small implementation
> details between the two.

> Right now I'm manually comparing the output of perf report to
> determine the salient differences.

> However, I'm wondering if there's a tool or otherwise best practice to
> essentially "diff" the two outputs.

> I could just use diff on the --stdio version of the reports, but
> sometimes lines get mangled and I loose context/hierarchy.

Can you try:

perf diff --help

- Arnaldo

Re: Comparing output of two perf cpu profiling reports

[Mark Davis]

Yeah, that would do it. Thanks to you both, Alex and Arnaldo!

On Thu, Apr 28, 2016 at 10:16 AM, Arnaldo Carvalho de Melo
<arnaldo.melo@gmail.com> wrote:
> Em Thu, Apr 28, 2016 at 09:27:07AM -0400, Mark Davis escreveu:
>> Hi, what is the recommend way to compare two different perf reports
>> from CPU event profiling (via sampling with perf record)?
>
>> I'm comparing two macro benchmarks, with a couple small implementation
>> details between the two.
>
>> Right now I'm manually comparing the output of perf report to
>> determine the salient differences.
>
>> However, I'm wondering if there's a tool or otherwise best practice to
>> essentially "diff" the two outputs.
>
>> I could just use diff on the --stdio version of the reports, but
>> sometimes lines get mangled and I loose context/hierarchy.
>
> Can you try:
>
> perf diff --help
>
> - Arnaldo