This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Serious performance regression -- some tree optimizer questions
- From: Ulrich Weigand <Ulrich dot Weigand at de dot ibm dot com>
- To: Andrew Pinski <pinskia at physics dot uc dot edu>
- Cc: gcc at gcc dot gnu dot org
- Date: Sat, 18 Dec 2004 00:50:14 +0100
- Subject: Re: Serious performance regression -- some tree optimizer questions
Andrew Pinski <pinskia@physics.uc.edu> wrote on 12/18/2004 12:23:09 AM:
> I should note that there were a recent patch to IVOPTs that was
> approved but not yet applied (why, I don't know) that should help.
> This patch helps us find the global minimum instead of a local
> minimum. Hopefully this patch helps this case also, it might not
> but who knows.
I already tried that one, but it doesn't help here. One problem
w.r.t. the loop optimizer with my problem is that there are too
many candidates, and the 'important' candidate heuristic doesn't
find the right ones. This can be circumvented by increasing the
iv-consider-all-candidates-bound parameter, which helps a lot
on 32-bit; on 64-bit there is the additional problem that the
optimizer cannot prove the 32-bit index arithmetic doesn't
overflow :-(
The other problem is the array issue I mentioned earlier.
Thanks,
Ulrich