diff options
| author | Craig Topper <craig.topper@gmail.com> | 2012-01-13 08:12:35 +0000 |
|---|---|---|
| committer | Craig Topper <craig.topper@gmail.com> | 2012-01-13 08:12:35 +0000 |
| commit | cb7e13d7c0f6b2c88020a21b0a0e1417f02969f6 (patch) | |
| tree | 3ff4bfab8a76df8ce48a7f2b36a788c87fade2ba /llvm/lib/Target/X86/X86MCInstLower.cpp | |
| parent | 1b845f712a407153f8399fbe62b302ba59fab67b (diff) | |
| download | bcm5719-llvm-cb7e13d7c0f6b2c88020a21b0a0e1417f02969f6.tar.gz bcm5719-llvm-cb7e13d7c0f6b2c88020a21b0a0e1417f02969f6.zip | |
Make X86 instruction selection use 256-bit VPXOR for build_vector of all ones if AVX2 is enabled. This gives the ExeDepsFix pass a chance to choose FP vs int as appropriate. Also use v8i32 as the type for getZeroVector if AVX2 is enabled. This is consistent with SSE2 using prefering v4i32.
llvm-svn: 148108
Diffstat (limited to 'llvm/lib/Target/X86/X86MCInstLower.cpp')
| -rw-r--r-- | llvm/lib/Target/X86/X86MCInstLower.cpp | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/llvm/lib/Target/X86/X86MCInstLower.cpp b/llvm/lib/Target/X86/X86MCInstLower.cpp index 92321961392..1f7d42240c1 100644 --- a/llvm/lib/Target/X86/X86MCInstLower.cpp +++ b/llvm/lib/Target/X86/X86MCInstLower.cpp @@ -373,6 +373,7 @@ ReSimplify: case X86::AVX_SET0PDY: LowerUnaryToTwoAddr(OutMI, X86::VXORPDYrr); break; case X86::AVX_SETALLONES: LowerUnaryToTwoAddr(OutMI, X86::VPCMPEQDrr); break; case X86::AVX2_SETALLONES: LowerUnaryToTwoAddr(OutMI, X86::VPCMPEQDYrr);break; + case X86::AVX2_SET0: LowerUnaryToTwoAddr(OutMI, X86::VPXORYrr); break; case X86::MOV16r0: LowerSubReg32_Op0(OutMI, X86::MOV32r0); // MOV16r0 -> MOV32r0 |
