This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH RS6000] peephole changing a move to itself with compare to a compare
- From: Michael Meissner <meissner at the-meissners dot org>
- To: Andrew Pinski <pinskia at physics dot uc dot edu>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Tue, 13 May 2003 12:17:52 -0400
- Subject: Re: [PATCH RS6000] peephole changing a move to itself with compare to a compare
- References: <6996E6E0-8553-11D7-ADD1-000393A6D2F2@physics.uc.edu>
On Tue, May 13, 2003 at 10:58:53AM -0400, Andrew Pinski wrote:
> This peephole changes `mr. rA, rA' to `cmp{w,d}i rA, 0' which is faster
> on
> the power 4, note I think there is a deficiency in a generic part of
> gcc's back-end but this was the fastest way and easiest way to fix it.
I am not responsible for the rs6000 backend any more, but I don't think a
peephole is the right approach for this patch. I would think modifying the
define_split to not do the move would be a better task.
--
Michael Meissner
email: meissner@the-meissners.org
http://www.the-meissners.org