<div dir="ltr"><div style="font-family:arial,sans-serif;font-size:13px"><div>I did profiling for </div><div>$python -m timeit -n 1000000 -s 'import numpy as np;x = np.asarray(1.0)' 'x+x'</div><div></div></div>
<div style="font-family:arial,sans-serif;font-size:13px">with oprofilier, and used gprof2dot.py to create callgraph, but I got graph[1] which doesn't create any meaning.<br></div><div style="font-family:arial,sans-serif;font-size:13px">
<br></div><div style="font-family:arial,sans-serif;font-size:13px">I tried to use pprof, but I can not find profiles to be used. like ls.prof in "pprof /bin/ls ls.prof"</div><div style="font-family:arial,sans-serif;font-size:13px">
Does any one has used pprof or any other profiler for generating callgraph?</div><div style="font-family:arial,sans-serif;font-size:13px"><br></div><div style="font-family:arial,sans-serif;font-size:13px">[1] <a href="https://docs.google.com/file/d/0B3Pqyp8kuQw0V042bmUwNHktZHM/edit" target="_blank">https://docs.google.com/file/d/0B3Pqyp8kuQw0V042bmUwNHktZHM/edit</a></div>
<div class="gmail_extra"><br><br><div class="gmail_quote">On Mon, May 6, 2013 at 4:56 AM, Nathaniel Smith <span dir="ltr"><<a href="mailto:njs@pobox.com" target="_blank">njs@pobox.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div><div>On Sun, May 5, 2013 at 5:57 PM, David Cournapeau <<a href="mailto:cournape@gmail.com" target="_blank">cournape@gmail.com</a>> wrote:<br>
>> perf is a fabulous framework and doesn't have any way to get full<br>
>> callgraph information out so IME it's been useless. They have<br>
>> reporting modes that claim to (like some "fractal" thing?) but AFAI<br>
>> been able to tell from docs/googling/mailing lists, there is nobody<br>
>> who understands how to interpret this output except the people who<br>
>> wrote it. Really a shame that it falls down in the last mile like<br>
>> that, hopefully they will fix this soon.<br>
><br>
> Perf doc is written for Vulcan, but it does what I think you want, say:<br>
><br>
> void work(int n) {<br>
> volatile int i=0; //don't optimize away<br>
> while(i++ < n);<br>
> }<br>
> void easy() { work(1000 * 1000 * 50); }<br>
> void hard() { work(1000*1000*1000); }<br>
> int main() { easy(); hard(); }<br>
><br>
> compile with gcc -g -O0, and then:<br>
><br>
> perf record -g -a -- ./a.out<br>
> perf report -g -a --stdio<br>
><br>
> gives me<br>
><br>
> 95.22% a.out a.out<br>
> [.] work<br>
> |<br>
> --- work<br>
> |<br>
> |--89.84%-- hard<br>
> | main<br>
> | __libc_start_main<br>
> |<br>
> --5.38%-- easy<br>
> main<br>
> __libc_start_main<br>
><br>
><br>
> or maybe even better with the -G option<br>
><br>
> 95.22% a.out a.out<br>
> [.] work<br>
> |<br>
> --- __libc_start_main<br>
> main<br>
> |<br>
> |--94.35%-- hard<br>
> | work<br>
> |<br>
> --5.65%-- easy<br>
> work<br>
><br>
<br>
</div></div>Yeah I've seen these displays before and I can see the information is<br>
there, and (knowing the code you ran) that somehow the first number<br>
has to do with the time spent under 'hard' and the second to do with<br>
time spent under 'easy', but I have no idea how to generalize this to<br>
arbitrary samples of these output formats. That's what I meant.<br>
<span><font color="#888888"><br>
-n<br>
</font></span><div><div>_______________________________________________<br>
NumPy-Discussion mailing list<br>
<a href="mailto:NumPy-Discussion@scipy.org" target="_blank">NumPy-Discussion@scipy.org</a><br>
<a href="http://mail.scipy.org/mailman/listinfo/numpy-discussion" target="_blank">http://mail.scipy.org/mailman/listinfo/numpy-discussion</a><br>
</div></div></blockquote></div><br><br clear="all"><div><br></div>-- <br><div dir="ltr"><div style="outline-style:none">Arink Verma</div><div style="outline-style:none">Indian Institute of Technology</div><div style="outline-style:none">
<a href="http://www.arinkverma.in">www.arinkverma.in</a></div><div style="outline-style:none"></div></div>
</div></div>