[LLVMbugs] [Bug 7518] [x86] terrible codegen of <2 x float>

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Tue Jun 29 10:51:52 CDT 2010


http://llvm.org/bugs/show_bug.cgi?id=7518

Dan Gohman <gohman at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |WONTFIX

--- Comment #3 from Dan Gohman <gohman at apple.com> 2010-06-29 10:51:51 CDT ---
This doesn't reproduce as is, but I think I know what's going on;
If I rewrite your <2 x float> testcase in C, I can reproduce it.

The fix is to use -disable-mmx (via -mllvm -disable-mmx in clang
and llvm-gcc). This option makes your LLVM IR <2 x float> case better,
and the C <2 x float> case better, and lots of other stuff better
besides.

Fixing codegen so that it generates optimal code automatically is
beyond the scope of this bug.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.


More information about the LLVMbugs mailing list