[llvm] r275770 - [X86] Add VPMULLW/D/Q instructions to X86InstrInfo::isAssociativeAndCommutative.
Craig Topper via llvm-commits
llvm-commits at lists.llvm.org
Sun Jul 17 23:14:57 PDT 2016
Author: ctopper
Date: Mon Jul 18 01:14:57 2016
New Revision: 275770
URL: http://llvm.org/viewvc/llvm-project?rev=275770&view=rev
Log:
[X86] Add VPMULLW/D/Q instructions to X86InstrInfo::isAssociativeAndCommutative.
Modified:
llvm/trunk/lib/Target/X86/X86InstrInfo.cpp
Modified: llvm/trunk/lib/Target/X86/X86InstrInfo.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86InstrInfo.cpp?rev=275770&r1=275769&r2=275770&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86InstrInfo.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86InstrInfo.cpp Mon Jul 18 01:14:57 2016
@@ -7727,6 +7727,19 @@ bool X86InstrInfo::isAssociativeAndCommu
case X86::VPADDWZrr:
case X86::VPADDDZrr:
case X86::VPADDQZrr:
+ case X86::VPMULLWrr:
+ case X86::VPMULLWYrr:
+ case X86::VPMULLWZ128rr:
+ case X86::VPMULLWZ256rr:
+ case X86::VPMULLWZrr:
+ case X86::VPMULLDrr:
+ case X86::VPMULLDYrr:
+ case X86::VPMULLDZ128rr:
+ case X86::VPMULLDZ256rr:
+ case X86::VPMULLDZrr:
+ case X86::VPMULLQZ128rr:
+ case X86::VPMULLQZ256rr:
+ case X86::VPMULLQZrr:
// Normal min/max instructions are not commutative because of NaN and signed
// zero semantics, but these are. Thus, there's no need to check for global
// relaxed math; the instructions themselves have the properties we need.
More information about the llvm-commits
mailing list