[llvm] r190258 - Using popcount should check the popcount feature flag not the SSE41 feature flag.
Craig Topper
craig.topper at gmail.com
Sat Sep 7 17:47:31 PDT 2013
Author: ctopper
Date: Sat Sep 7 19:47:31 2013
New Revision: 190258
URL: http://llvm.org/viewvc/llvm-project?rev=190258&view=rev
Log:
Using popcount should check the popcount feature flag not the SSE41 feature flag.
Modified:
llvm/trunk/lib/Target/X86/X86TargetTransformInfo.cpp
Modified: llvm/trunk/lib/Target/X86/X86TargetTransformInfo.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86TargetTransformInfo.cpp?rev=190258&r1=190257&r2=190258&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86TargetTransformInfo.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86TargetTransformInfo.cpp Sat Sep 7 19:47:31 2013
@@ -127,8 +127,8 @@ X86TTI::PopcntSupportKind X86TTI::getPop
assert(isPowerOf2_32(TyWidth) && "Ty width must be power of 2");
// TODO: Currently the __builtin_popcount() implementation using SSE3
// instructions is inefficient. Once the problem is fixed, we should
- // call ST->hasSSE3() instead of ST->hasSSE4().
- return ST->hasSSE41() ? PSK_FastHardware : PSK_Software;
+ // call ST->hasSSE3() instead of ST->hasPOPCNT().
+ return ST->hasPOPCNT() ? PSK_FastHardware : PSK_Software;
}
unsigned X86TTI::getNumberOfRegisters(bool Vector) const {
More information about the llvm-commits
mailing list