[LLVMbugs] [Bug 10580] [AVX] vector comparisons are scalarized

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Mon Aug 8 17:49:31 PDT 2011


Bruno Cardoso Lopes <bruno.cardoso at gmail.com> changed:

           What    |Removed                     |Added
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #1 from Bruno Cardoso Lopes <bruno.cardoso at gmail.com> 2011-08-08 19:49:30 CDT ---
Hi Matt,

I've added the AVX part of this in r137090. But keep in mind that if you want
to generate cmp*ps/pd variants you should sign extended the result, so in the
correct testcase is:

define <8 x i32> @cmp00(<8 x float> %a, <8 x float> %b) nounwind readnone {
  %bincmp = fcmp olt <8 x float> %a, %b
  %s = sext <8 x i1> %bincmp to <8 x i32>
  ret <8 x i32> %s

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 llvm-bugs mailing list