This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
patch/RFC: PR optimization/8423
- From: Roger Sayle <roger at eyesopen dot com>
- To: Zack Weinberg <zack at codesourcery dot com>
- Cc: <gcc-patches at gcc dot gnu dot org>
- Date: Tue, 5 Nov 2002 17:03:00 -0700 (MST)
- Subject: patch/RFC: PR optimization/8423
Could you confirm that this also solves GNATS PR middle-end/3919
which I believe is the same failure?
I'd feel much happier if we had some benchmarking results for such
a major change to GCC's pass ordering. The proposed change could
have catastrophic effects on GCC's performance for an extension
that shouldn't be relied upon to recognize all possible invariants.
[See RTH's comments in 3919's audit trail!]
I'm not against the patch per se, but perhaps a safer change would
be to introduce a pass or global variable indicating when its safe
to optimize __builtin_constant_p as zero. i.e. not until GCSE.
Roger
--