This is the mail archive of the
mailing list for the GCC project.
Re: gcc compile-time performance
- From: Andi Kleen <ak at suse dot de>
- To: gcc at gcc dot gnu dot org, dhazeghi at pacbell dot net, neil at daikokuya dot demon dot co dot uk
- Date: 16 May 2002 16:34:32 +0200
- Subject: Re: gcc compile-time performance
- References: <firstname.lastname@example.org>
Dara Hazeghi <email@example.com> writes:
> Obiously you are free to draw your own conclusions ;-) Now far be it for me
> (who haven't contributed a line of code) to criticize, but at least with
> respect to the optimized build, the slowdown from 3.0 to 3.1 is considerable
> (a 38% increase in compile time). So some food for thought...
I guess the only way to avoid such things in the future would be to track
the compiler performance of mainline (similar to how the performance of the
resulting code is tracked at http://www.suse.de/~aj/SPEC/index.html).
Then it would be obvious which changes caused bad performance regressions.
Another way would be to do regular profiling and publishing of the worst
"offending" functions. With an global system
profile like oprofile for linux (http://oprofile.sourceforge.net) it is
possible to e.g. profile a complete bootstrap, not just individual runs,
which should give a quite good general picture of the performance
problems in the profiler.