diff options
| author | Craig Topper <craig.topper@gmail.com> | 2012-01-13 06:59:47 +0000 |
|---|---|---|
| committer | Craig Topper <craig.topper@gmail.com> | 2012-01-13 06:59:47 +0000 |
| commit | 9f14d9f9392d46f81722c8342b149e0051ff91cc (patch) | |
| tree | 9053706a75cad8382ed811c44b32762c25895539 /llvm | |
| parent | e77e84e4b7cda4607e053c3521e61ce9a961a847 (diff) | |
| download | bcm5719-llvm-9f14d9f9392d46f81722c8342b149e0051ff91cc.tar.gz bcm5719-llvm-9f14d9f9392d46f81722c8342b149e0051ff91cc.zip | |
Add patterns for v16i16 and v32i8 immAllZerosV to select VPXOR to match v4i64 and v8i32.
llvm-svn: 148106
Diffstat (limited to 'llvm')
| -rw-r--r-- | llvm/lib/Target/X86/X86InstrSSE.td | 8 | ||||
| -rw-r--r-- | llvm/test/CodeGen/X86/avx2-shift.ll | 2 |
2 files changed, 9 insertions, 1 deletions
diff --git a/llvm/lib/Target/X86/X86InstrSSE.td b/llvm/lib/Target/X86/X86InstrSSE.td index 51cc7f384e5..d3cd6a82fdb 100644 --- a/llvm/lib/Target/X86/X86InstrSSE.td +++ b/llvm/lib/Target/X86/X86InstrSSE.td @@ -289,6 +289,14 @@ def AVX_SET0PDY : PDI<0x57, MRMInitReg, (outs VR256:$dst), (ins), "", // AVX has no support for 256-bit integer instructions, but since the 128-bit // VPXOR instruction writes zero to its upper part, it's safe build zeros. +def : Pat<(v32i8 immAllZerosV), (SUBREG_TO_REG (i8 0), (V_SET0), sub_xmm)>; +def : Pat<(bc_v32i8 (v8f32 immAllZerosV)), + (SUBREG_TO_REG (i8 0), (V_SET0), sub_xmm)>; + +def : Pat<(v16i16 immAllZerosV), (SUBREG_TO_REG (i16 0), (V_SET0), sub_xmm)>; +def : Pat<(bc_v16i16 (v8f32 immAllZerosV)), + (SUBREG_TO_REG (i16 0), (V_SET0), sub_xmm)>; + def : Pat<(v8i32 immAllZerosV), (SUBREG_TO_REG (i32 0), (V_SET0), sub_xmm)>; def : Pat<(bc_v8i32 (v8f32 immAllZerosV)), (SUBREG_TO_REG (i32 0), (V_SET0), sub_xmm)>; diff --git a/llvm/test/CodeGen/X86/avx2-shift.ll b/llvm/test/CodeGen/X86/avx2-shift.ll index b6cf54ebe8f..1f192a0d323 100644 --- a/llvm/test/CodeGen/X86/avx2-shift.ll +++ b/llvm/test/CodeGen/X86/avx2-shift.ll @@ -231,7 +231,7 @@ define <32 x i8> @sra_v32i8_7(<32 x i8> %A) nounwind { %B = ashr <32 x i8> %A, <i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7> ret <32 x i8> %B ; CHECK: sra_v32i8_7: -; CHECK: vxorps +; CHECK: vpxor ; CHECK: vpcmpgtb ; CHECK: ret } |
