This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix PR58682
- From: Jan Hubicka <hubicka at ucw dot cz>
- To: Paulo Matos <pmatos at broadcom dot com>
- Cc: Jan Hubicka <hubicka at ucw dot cz>, Richard Biener <richard dot guenther at gmail dot com>, Mike Stump <mikestump at comcast dot net>, Kyrill Tkachov <kyrylo dot tkachov at arm dot com>, "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>
- Date: Mon, 21 Oct 2013 17:18:46 +0200
- Subject: Re: [PATCH] Fix PR58682
- Authentication-results: sourceware.org; auth=none
- References: <19EB96622A777C4AB91610E763265F462DB4E2 at SJEXCHMB14 dot corp dot ad dot broadcom dot com> <19EB96622A777C4AB91610E763265F462DB4F8 at SJEXCHMB14 dot corp dot ad dot broadcom dot com> <525BCADE dot 8040306 at arm dot com> <EB210B59-0645-4C3D-B7DD-56B4A64F6A22 at comcast dot net> <19EB96622A777C4AB91610E763265F462DEAC2 at SJEXCHMB14 dot corp dot ad dot broadcom dot com> <19EB96622A777C4AB91610E763265F462E18DA at SJEXCHMB14 dot corp dot ad dot broadcom dot com> <CAFiYyc2z_++BKiOD9OHDSQ603tBB9mOuWpoRZVNLHQEL4B_cxw at mail dot gmail dot com> <19EB96622A777C4AB91610E763265F462E31CD at SJEXCHMB14 dot corp dot ad dot broadcom dot com> <20131021122122 dot GA21513 at kam dot mff dot cuni dot cz> <19EB96622A777C4AB91610E763265F462E3341 at SJEXCHMB14 dot corp dot ad dot broadcom dot com>
> > -----Original Message-----
> > From: Jan Hubicka [mailto:hubicka@ucw.cz]
> > Sent: 21 October 2013 13:21
> > To: Paulo Matos
> > Cc: Richard Biener; Mike Stump; Kyrill Tkachov; gcc-patches@gcc.gnu.org; Jan
> > Hubicka (hubicka@ucw.cz)
> > Subject: Re: [PATCH] Fix PR58682
> >
> > This won't work - the max_count is there to get things correctly scaled by
> > badness calculation.
> > Changing max_count would mean the need to recompute all badnesses in the fibheap
> > keys.
> > I would just cap the edge_count in badness calculation.
> >
> > Honza
> >
>
> Makes sense. Thanks for the comment. I have created a new patch. Now attached.
>
> 2013-10-21 Paulo Matos <pmatos@broadcom.com>
>
> * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
> calculations.
OK, thanks!
Honza