diff options
Diffstat (limited to 'llvm/test/CodeGen/X86/combine-pavg.ll')
| -rw-r--r-- | llvm/test/CodeGen/X86/combine-pavg.ll | 34 |
1 files changed, 32 insertions, 2 deletions
diff --git a/llvm/test/CodeGen/X86/combine-pavg.ll b/llvm/test/CodeGen/X86/combine-pavg.ll index c6e1c242565b..bfeebb473c74 100644 --- a/llvm/test/CodeGen/X86/combine-pavg.ll +++ b/llvm/test/CodeGen/X86/combine-pavg.ll @@ -1,7 +1,7 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py ; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=SSE -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX -; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=AVX,AVX1 +; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX2 declare <16 x i8> @llvm.x86.sse2.pavg.b(<16 x i8>, <16 x i8>) nounwind readnone declare <8 x i16> @llvm.x86.sse2.pavg.w(<8 x i16>, <8 x i16>) nounwind readnone @@ -37,6 +37,36 @@ define <16 x i8> @combine_pavgw_knownbits(<8 x i16> %a0, <8 x i16> %a1, <8 x i16 ; SSE-NEXT: pand %xmm1, %xmm0 ; SSE-NEXT: packuswb %xmm3, %xmm0 ; SSE-NEXT: retq +; +; AVX1-LABEL: combine_pavgw_knownbits: +; AVX1: # %bb.0: +; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [31,31,31,31,31,31,31,31] +; AVX1-NEXT: vpand %xmm4, %xmm0, %xmm0 +; AVX1-NEXT: vpand %xmm4, %xmm1, %xmm1 +; AVX1-NEXT: vpavgw %xmm1, %xmm0, %xmm0 +; AVX1-NEXT: vpand %xmm4, %xmm2, %xmm1 +; AVX1-NEXT: vpand %xmm4, %xmm3, %xmm2 +; AVX1-NEXT: vpavgw %xmm2, %xmm1, %xmm1 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [255,0,255,0,255,0,255,0,255,0,255,0,255,0,255,0] +; AVX1-NEXT: vpand %xmm2, %xmm1, %xmm1 +; AVX1-NEXT: vpand %xmm2, %xmm0, %xmm0 +; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0 +; AVX1-NEXT: retq +; +; AVX2-LABEL: combine_pavgw_knownbits: +; AVX2: # %bb.0: +; AVX2-NEXT: vmovdqa {{.*#+}} xmm4 = [31,31,31,31,31,31,31,31] +; AVX2-NEXT: vpand %xmm4, %xmm0, %xmm0 +; AVX2-NEXT: vpand %xmm4, %xmm1, %xmm1 +; AVX2-NEXT: vpavgw %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: vpand %xmm4, %xmm2, %xmm1 +; AVX2-NEXT: vpand %xmm4, %xmm3, %xmm2 +; AVX2-NEXT: vpavgw %xmm2, %xmm1, %xmm1 +; AVX2-NEXT: vpbroadcastw {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255] +; AVX2-NEXT: vpand %xmm2, %xmm1, %xmm1 +; AVX2-NEXT: vpand %xmm2, %xmm0, %xmm0 +; AVX2-NEXT: vpackuswb %xmm1, %xmm0, %xmm0 +; AVX2-NEXT: retq %m0 = and <8 x i16> %a0, <i16 31, i16 31, i16 31, i16 31, i16 31, i16 31, i16 31, i16 31> %m1 = and <8 x i16> %a1, <i16 31, i16 31, i16 31, i16 31, i16 31, i16 31, i16 31, i16 31> %m2 = and <8 x i16> %a2, <i16 31, i16 31, i16 31, i16 31, i16 31, i16 31, i16 31, i16 31> |