This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug c++/11376] [3.3/3.4 regression] mozilla-1.4 miscompiled
- From: "pinskia at physics dot uc dot edu" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 8 Jul 2003 19:59:05 -0000
- Subject: [Bug c++/11376] [3.3/3.4 regression] mozilla-1.4 miscompiled
- References: <20030630081619.11376.sirl@gcc.gnu.org>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11376
------- Additional Comments From pinskia at physics dot uc dot edu 2003-07-08 19:59 -------
replying to comment 24:
The reason why x86 does not have this problem is because the schedular does not move around
load and stores as much as the PPC schedular does, look at the gcc which is SPEC, it contains non
alias safe code and it passes comparison test using the pentium 3 schedular but not with the G4
PPC one, it fails. Another reason why it effects PPC linux only and not PPC darwin is that the PIC
code generation is different so you get different code for the PIC code, it might effect PPC darwin
(in fact the simplified testcase does seg fault on PPC darwin). So the scheduling of the asm causes
the load at different times.