This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Performance regression due to combine.c patch
- From: Kazu Hirata <kazu at cs dot umass dot edu>
- To: Ulrich dot Weigand at de dot ibm dot com
- Cc: rth at redhat dot com, gcc-patches at gcc dot gnu dot org
- Date: Wed, 17 Dec 2003 21:21:38 -0500 (EST)
- Subject: Re: Performance regression due to combine.c patch
- References: <OF88A6957D.2AB80DAD-ONC1256DFF.00468E4F@de.ibm.com>
Hi Ulrich,
> Why is this a problem? Paradoxical subregs of a pseudo used as SET_DEST
> should be just fine as far as I can see ...
If I allow a paradoxical subreg there, I get an ICE in local-alloc.c.
(I just reverted the patch in my local tree and confirmed the
regressions that I used to see before the patch.) Maybe it's time to
look at the real cause of this problem. I am not familiar with
local-alloc.c, so don't hold your breath.
Kazu Hirata