; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+ssse3,+sse4a | FileCheck %s --check-prefixes=CHECK,SSSE3 ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse4.2,+sse4a | FileCheck %s --check-prefixes=CHECK,SSE42 ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx,+sse4a| FileCheck %s --check-prefixes=CHECK,AVX ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx2,+sse4a | FileCheck %s --check-prefixes=CHECK,AVX ; ; Combine tests involving SSE4A target shuffles (EXTRQI,INSERTQI) declare <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8>, <16 x i8>) define <16 x i8> @combine_extrqi_pshufb_16i8(<16 x i8> %a0) { ; CHECK-LABEL: combine_extrqi_pshufb_16i8: ; CHECK: # %bb.0: ; CHECK-NEXT: extrq {{.*#+}} xmm0 = xmm0[1,2],zero,zero,zero,zero,zero,zero,xmm0[u,u,u,u,u,u,u,u] ; CHECK-NEXT: retq %1 = shufflevector <16 x i8> %a0, <16 x i8> zeroinitializer, <16 x i32> %2 = tail call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %1, <16 x i8> ) ret <16 x i8> %2 } define <8 x i16> @combine_extrqi_pshufb_8i16(<8 x i16> %a0) { ; CHECK-LABEL: combine_extrqi_pshufb_8i16: ; CHECK: # %bb.0: ; CHECK-NEXT: extrq {{.*#+}} xmm0 = xmm0[2,3],zero,zero,zero,zero,zero,zero,xmm0[u,u,u,u,u,u,u,u] ; CHECK-NEXT: retq %1 = shufflevector <8 x i16> %a0, <8 x i16> zeroinitializer, <8 x i32> %2 = bitcast <8 x i16> %1 to <16 x i8> %3 = tail call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %2, <16 x i8> ) %4 = bitcast <16 x i8> %3 to <8 x i16> ret <8 x i16> %4 } define <16 x i8> @combine_insertqi_pshufb_16i8(<16 x i8> %a0, <16 x i8> %a1) { ; SSSE3-LABEL: combine_insertqi_pshufb_16i8: ; SSSE3: # %bb.0: ; SSSE3-NEXT: movdqa %xmm1, %xmm0 ; SSSE3-NEXT: extrq {{.*#+}} xmm0 = xmm0[0,1],zero,zero,zero,zero,zero,zero,xmm0[u,u,u,u,u,u,u,u] ; SSSE3-NEXT: retq ; ; SSE42-LABEL: combine_insertqi_pshufb_16i8: ; SSE42: # %bb.0: ; SSE42-NEXT: pmovzxwq {{.*#+}} xmm0 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero ; SSE42-NEXT: retq ; ; AVX-LABEL: combine_insertqi_pshufb_16i8: ; AVX: # %bb.0: ; AVX-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero ; AVX-NEXT: retq %1 = shufflevector <16 x i8> %a0, <16 x i8> %a1, <16 x i32> %2 = tail call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %1, <16 x i8> ) ret <16 x i8> %2 } define <8 x i16> @combine_insertqi_pshufb_8i16(<8 x i16> %a0, <8 x i16> %a1) { ; SSSE3-LABEL: combine_insertqi_pshufb_8i16: ; SSSE3: # %bb.0: ; SSSE3-NEXT: movdqa %xmm1, %xmm0 ; SSSE3-NEXT: extrq {{.*#+}} xmm0 = xmm0[0,1],zero,zero,zero,zero,zero,zero,xmm0[u,u,u,u,u,u,u,u] ; SSSE3-NEXT: retq ; ; SSE42-LABEL: combine_insertqi_pshufb_8i16: ; SSE42: # %bb.0: ; SSE42-NEXT: pmovzxwq {{.*#+}} xmm0 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero ; SSE42-NEXT: retq ; ; AVX-LABEL: combine_insertqi_pshufb_8i16: ; AVX: # %bb.0: ; AVX-NEXT: vpmovzxwq {{.*#+}} xmm0 = xmm1[0],zero,zero,zero,xmm1[1],zero,zero,zero ; AVX-NEXT: retq %1 = shufflevector <8 x i16> %a0, <8 x i16> %a1, <8 x i32> %2 = bitcast <8 x i16> %1 to <16 x i8> %3 = tail call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %2, <16 x i8> ) %4 = bitcast <16 x i8> %3 to <8 x i16> ret <8 x i16> %4 } define <16 x i8> @combine_pshufb_insertqi_pshufb(<16 x i8> %a0, <16 x i8> %a1) { ; CHECK-LABEL: combine_pshufb_insertqi_pshufb: ; CHECK: # %bb.0: ; CHECK-NEXT: insertq {{.*#+}} xmm0 = xmm0[0],xmm1[0,1],xmm0[3,4,5,6,7,u,u,u,u,u,u,u,u] ; CHECK-NEXT: retq %1 = tail call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %a0, <16 x i8> ) %2 = shufflevector <16 x i8> %1, <16 x i8> %a1, <16 x i32> %3 = tail call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %2, <16 x i8> ) ret <16 x i8> %3 }