This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug target/33135] [SH] -ffinite-math-only should not be on by default
- From: "olegendo at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Mon, 16 Jul 2012 20:50:50 +0000
- Subject: [Bug target/33135] [SH] -ffinite-math-only should not be on by default
- Auto-submitted: auto-generated
- References: <bug-33135-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33135
--- Comment #4 from Oleg Endo <olegendo at gcc dot gnu.org> 2012-07-16 20:50:50 UTC ---
(In reply to comment #3)
> I guess that the better default would be -mno-ieee for bare metals
Hm ... Is there a particular reason to distinguish between linux and non-linux?
I was actually thinking of tying -mno-ieee to -ffinite-math-only in the
following way:
no options -> implicit -fno-finite-math-only -mieee
-ffinite-math-only -> disable ieee fcmp, allow other optimizations
-ffinite-math-only -mieee -> enable ieee fcmp, allow other optimizations
-mno-ieee -> disable ieee fcmp, disallow other optimizations
I think if faster FPU code is desired, most likely the -ffast-math (or some
sub-option like -ffinite-math-only) will be consulted in the first place. In
case of bare metal configurations, users will probably tweak their specific
options.
Thus, I did not want to distinguish between linux or non-linux here.