This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug middle-end/31862] Loop IM and other optimizations harmful for -fopenmp
- From: "theodore dot papadopoulo at sophia dot inria dot fr" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 4 Jun 2007 18:12:33 -0000
- Subject: [Bug middle-end/31862] Loop IM and other optimizations harmful for -fopenmp
- References: <bug-31862-87@http.gcc.gnu.org/bugzilla/>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Comment #11 from theodore dot papadopoulo at sophia dot inria dot fr 2007-06-04 18:12 -------
(In reply to comment #8)
I suspect (unless I'm overlooked something) that this problem cause wrong
statistics when using jointly the -fopenmp and -profile-* flags.
I tried that and seen messages saying that the counters are corrupted (negative
counters)... Still need to find a small testcase though.
If I'm true, at the very least, the flags should error when used in conjunction
(for this I can do a patch, if this is the right solution).
Correcting this in this case should be easy (multiplicate the counters by the
number of threads and join them at dump time), but that's probably beyond my
gcc hacking capabilities...
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31862