1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py 2; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512vpopcntdq,+prefer-256-bit | FileCheck %s --check-prefix=CHECK --check-prefix=AVX256 3; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512vpopcntdq,-prefer-256-bit | FileCheck %s --check-prefix=CHECK --check-prefix=AVX512 --check-prefix=AVX512VL 4; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vpopcntdq,+prefer-256-bit | FileCheck %s --check-prefix=CHECK --check-prefix=AVX512 --check-prefix=AVX512F 5; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f,+avx512vpopcntdq,-prefer-256-bit | FileCheck %s --check-prefix=CHECK --check-prefix=AVX512 --check-prefix=AVX512F 6 7define <8 x i16> @testv8i16(<8 x i16> %in) { 8; AVX256-LABEL: testv8i16: 9; AVX256: # %bb.0: 10; AVX256-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero 11; AVX256-NEXT: vpopcntd %ymm0, %ymm0 12; AVX256-NEXT: vpmovdw %ymm0, %xmm0 13; AVX256-NEXT: vzeroupper 14; AVX256-NEXT: retq 15; 16; AVX512VL-LABEL: testv8i16: 17; AVX512VL: # %bb.0: 18; AVX512VL-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero 19; AVX512VL-NEXT: vpopcntd %ymm0, %ymm0 20; AVX512VL-NEXT: vpmovdw %ymm0, %xmm0 21; AVX512VL-NEXT: vzeroupper 22; AVX512VL-NEXT: retq 23; 24; AVX512F-LABEL: testv8i16: 25; AVX512F: # %bb.0: 26; AVX512F-NEXT: vpmovzxwd {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero 27; AVX512F-NEXT: vpopcntd %zmm0, %zmm0 28; AVX512F-NEXT: vpmovdw %zmm0, %ymm0 29; AVX512F-NEXT: # kill: def $xmm0 killed $xmm0 killed $ymm0 30; AVX512F-NEXT: vzeroupper 31; AVX512F-NEXT: retq 32 %out = call <8 x i16> @llvm.ctpop.v8i16(<8 x i16> %in) 33 ret <8 x i16> %out 34} 35 36define <16 x i8> @testv16i8(<16 x i8> %in) { 37; AVX256-LABEL: testv16i8: 38; AVX256: # %bb.0: 39; AVX256-NEXT: vmovdqa {{.*#+}} xmm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15] 40; AVX256-NEXT: vpand %xmm1, %xmm0, %xmm2 41; AVX256-NEXT: vmovdqa {{.*#+}} xmm3 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4] 42; AVX256-NEXT: vpshufb %xmm2, %xmm3, %xmm2 43; AVX256-NEXT: vpsrlw $4, %xmm0, %xmm0 44; AVX256-NEXT: vpand %xmm1, %xmm0, %xmm0 45; AVX256-NEXT: vpshufb %xmm0, %xmm3, %xmm0 46; AVX256-NEXT: vpaddb %xmm2, %xmm0, %xmm0 47; AVX256-NEXT: retq 48; 49; AVX512-LABEL: testv16i8: 50; AVX512: # %bb.0: 51; AVX512-NEXT: vpmovzxbd {{.*#+}} zmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[9],zero,zero,zero,xmm0[10],zero,zero,zero,xmm0[11],zero,zero,zero,xmm0[12],zero,zero,zero,xmm0[13],zero,zero,zero,xmm0[14],zero,zero,zero,xmm0[15],zero,zero,zero 52; AVX512-NEXT: vpopcntd %zmm0, %zmm0 53; AVX512-NEXT: vpmovdb %zmm0, %xmm0 54; AVX512-NEXT: vzeroupper 55; AVX512-NEXT: retq 56 %out = call <16 x i8> @llvm.ctpop.v16i8(<16 x i8> %in) 57 ret <16 x i8> %out 58} 59 60define <16 x i16> @testv16i16(<16 x i16> %in) { 61; AVX256-LABEL: testv16i16: 62; AVX256: # %bb.0: 63; AVX256-NEXT: vmovdqa {{.*#+}} ymm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15] 64; AVX256-NEXT: vpand %ymm1, %ymm0, %ymm2 65; AVX256-NEXT: vmovdqa {{.*#+}} ymm3 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4,0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4] 66; AVX256-NEXT: vpshufb %ymm2, %ymm3, %ymm2 67; AVX256-NEXT: vpsrlw $4, %ymm0, %ymm0 68; AVX256-NEXT: vpand %ymm1, %ymm0, %ymm0 69; AVX256-NEXT: vpshufb %ymm0, %ymm3, %ymm0 70; AVX256-NEXT: vpaddb %ymm2, %ymm0, %ymm0 71; AVX256-NEXT: vpsllw $8, %ymm0, %ymm1 72; AVX256-NEXT: vpaddb %ymm0, %ymm1, %ymm0 73; AVX256-NEXT: vpsrlw $8, %ymm0, %ymm0 74; AVX256-NEXT: retq 75; 76; AVX512-LABEL: testv16i16: 77; AVX512: # %bb.0: 78; AVX512-NEXT: vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero 79; AVX512-NEXT: vpopcntd %zmm0, %zmm0 80; AVX512-NEXT: vpmovdw %zmm0, %ymm0 81; AVX512-NEXT: retq 82 %out = call <16 x i16> @llvm.ctpop.v16i16(<16 x i16> %in) 83 ret <16 x i16> %out 84} 85 86define <32 x i8> @testv32i8(<32 x i8> %in) { 87; CHECK-LABEL: testv32i8: 88; CHECK: # %bb.0: 89; CHECK-NEXT: vmovdqa {{.*#+}} ymm1 = [15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15,15] 90; CHECK-NEXT: vpand %ymm1, %ymm0, %ymm2 91; CHECK-NEXT: vmovdqa {{.*#+}} ymm3 = [0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4,0,1,1,2,1,2,2,3,1,2,2,3,2,3,3,4] 92; CHECK-NEXT: vpshufb %ymm2, %ymm3, %ymm2 93; CHECK-NEXT: vpsrlw $4, %ymm0, %ymm0 94; CHECK-NEXT: vpand %ymm1, %ymm0, %ymm0 95; CHECK-NEXT: vpshufb %ymm0, %ymm3, %ymm0 96; CHECK-NEXT: vpaddb %ymm2, %ymm0, %ymm0 97; CHECK-NEXT: retq 98 %out = call <32 x i8> @llvm.ctpop.v32i8(<32 x i8> %in) 99 ret <32 x i8> %out 100} 101 102declare <8 x i16> @llvm.ctpop.v8i16(<8 x i16>) 103declare <16 x i8> @llvm.ctpop.v16i8(<16 x i8>) 104declare <16 x i16> @llvm.ctpop.v16i16(<16 x i16>) 105declare <32 x i8> @llvm.ctpop.v32i8(<32 x i8>) 106