This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: [RFC] Improve Tree-SSA if-conversion - convergence of efforts
- From: Michael Matz <matz at suse dot de>
- To: Daniel Berlin <dberlin at dberlin dot org>
- Cc: Tehila Meyzels <TEHILA at il dot ibm dot com>, gcc-patches at gcc dot gnu dot org, gcc at gcc dot gnu dot org, trevor_smigiel at playstation dot sony dot com, Revital1 Eres <ERES at il dot ibm dot com>, Ulrich Weigand <Ulrich dot Weigand at de dot ibm dot com>, Victor Kaplansky <VICTORK at il dot ibm dot com>, dpatel at apple dot com
- Date: Tue, 31 Jul 2007 17:05:53 +0200 (CEST)
- Subject: Re: [RFC] Improve Tree-SSA if-conversion - convergence of efforts
- References: <OFE09A65AE.6EB97FFD-ONC2257329.004AA175-C2257329.004B47FB@il.ibm.com> <4aca3dc20707310800j41fb5cel4ef815bbad42be07@mail.gmail.com>
Hi,
On Tue, 31 Jul 2007, Daniel Berlin wrote:
> > 2. Store-sinking/load hoisting may have an overhead and may degrade
> > performance unless the relevant conditional branch gets if-converted.
>
> I agree with you for conditional stores/loads.
>
> The unconditional store/load stuff, however, is exactly what
> tree-ssa-sink was meant to do, and belongs there (this is #3 above). I'm
> certainly going to fight tooth and nail against trying to shoehorn
> unconditional store sinking into if-conv.
FWIW I also agree that handling unconditional stores/loads does not belong
in if-conv (or phi-opt), but in ssa-sink (or some similar transformation
which can or can not use value numbers and the like).
Ciao,
Michael.