This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: gcc compile-time performance
- From: Daniel Berlin <dberlin at dberlin dot org>
- To: Dara Hazeghi <dhazeghi at pacbell dot net>
- Cc: Neil Booth <neil at daikokuya dot demon dot co dot uk>, Andi Kleen <ak at suse dot de>,<gcc at gcc dot gnu dot org>
- Date: Fri, 17 May 2002 12:01:33 -0400 (EDT)
- Subject: Re: gcc compile-time performance
On Fri, 17 May 2002, Dara Hazeghi wrote:
> On Thursday 16 May 2002 12:09 pm, Neil Booth wrote:
> > Andi Kleen wrote:-
> >
> > > 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.
> >
> > I suspect that might be a good idea.
>
> Hmm...
>
> I think the next problem would be to determine what is and isn't a relevant
> test. Obviously gcc 2.7.2.3 works for testing the C front-end (heck, the SPEC
> folks agree too...). Would it make sense to come up with a bunch of
> "representative" applications for the various front-ends? Alternatively,
> would it make more sense to track the compile time of the SPEC builds
> (shouldn't be too difficult).
>
> Now the other problem is how to distinguish noise from real compile-time
> performance regressions, as much of the differences in performance up 'til
> now seem to have most likely been a cumulative effect of hundreds of patches
> (I have no data to back this up, just a personal opinion).
I think we actually *do* have data to back this up somewhere.
I also remember Stan Shebs mentioning it at some point (Stan, maybe i'm
misremembering, so if i just attributed something to you that you never
said, ....).
Much like a software project gets a year late one day at a time, gcc has
gotten 5 minutes slower at compiling one second at a time.
>
> Just thinking out loud...
>
> Thanks,
>
> Dara
>