This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [Patch] Splitting memory references during unrolling (resubmission)
- From: Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>
- To: Revital Eres <ERES at il dot ibm dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Sat, 21 Aug 2004 17:39:22 +0200
- Subject: Re: [Patch] Splitting memory references during unrolling (resubmission)
- References: <OF414523BA.9F1F83DE-ONC2256EEB.00205E24-C2256EEB.0023CD3E@il.ibm.com>
Hello,
> Following the fix for the unroller patch
> (http://gcc.gnu.org/ml/gcc-patches/2004-08/msg00397.html)
> I resubmit the patch for Splitting memory references during unrolling.
>
> Bootstrapped (with -funroll-all-loops) & regression tests on POWER4.
could you please measure the effect of this patch on some benchmark
on i686 and post results to the mailing list?
I suspect this optimization might spoil the code there; if this
turned out to be the case, you would have to restrict it to work
only on architectures where it is profitable.
Zdenek