This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug rtl-optimization/48823] gcc-4.6.0 floating-point optimization regression on ia64-Linux
- From: "jdemeyer at cage dot ugent.be" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Thu, 13 Oct 2011 20:05:21 +0000
- Subject: [Bug rtl-optimization/48823] gcc-4.6.0 floating-point optimization regression on ia64-Linux
- Auto-submitted: auto-generated
- References: <bug-48823-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48823
--- Comment #4 from Jeroen Demeyer <jdemeyer at cage dot ugent.be> 2011-10-13 20:05:21 UTC ---
(In reply to comment #3)
> The difference in results is due to the use of the fma instructions (fused
> multiply add). If you don't want the fma instructions to be used you should
> use the -ffp-contract=on flag.
Thanks for the reply, I understand your reasoning. However, will this option
guarantee strict IEEE floating-point behaviour or are there other flags to be
specified? I have always been under the impression that *by default* gcc
produced code respecting the floating-point standard unless one specifies
something like -ffast-math. It seems this has changed in recent GCC versions.