This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug tree-optimization/20643] Tree loop optimizer does worse job than RTL loop optimizer
- From: "dberlin at dberlin dot org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 26 Mar 2005 00:03:01 -0000
- Subject: [Bug tree-optimization/20643] Tree loop optimizer does worse job than RTL loop optimizer
- References: <20050325222146.20643.sje@cup.hp.com>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Additional Comments From dberlin at gcc dot gnu dot org 2005-03-26 00:02 -------
Subject: Re: New: Tree loop optimizer does
worse job than RTL loop optimizer
On Fri, 2005-03-25 at 22:21 +0000, sje at cup dot hp dot com wrote:
> In the attached test case, 3.4.* GCC generates better code than 4.0 (or 4.1)
> because it moves more loop invariant code out of the inner loop of P7Viterbi.
> The problem seems to be in the alias analysis which determines what can be moved
> out of that loop. If you change the field M, which is unused, from int to float
> then the 4.* GCC generates better code. I tried the structure-alias branch to
> see if that helped and it didn't.
Structure aliasing can't disambiguate this yet. It will soon enough.
I haven't been putting the remainder of the part2 work on the
struct-aliasing branch because it's all of one file and a very small
amount of integration work, and it wasn't worth keeping merges up to
date for it.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20643