1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py 2; RUN: llc < %s -mcpu=x86-64 -mattr=+avx512f | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512F 3; RUN: llc < %s -mcpu=x86-64 -mattr=+avx512bw | FileCheck %s --check-prefix=ALL --check-prefix=AVX512 --check-prefix=AVX512BW 4 5target triple = "x86_64-unknown-unknown" 6 7define <16 x float> @shuffle_v16f32_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<16 x float> %a, <16 x float> %b) { 8; ALL-LABEL: shuffle_v16f32_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00: 9; ALL: # BB#0: 10; ALL-NEXT: vbroadcastss %xmm0, %zmm0 11; ALL-NEXT: retq 12 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32><i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0> 13 ret <16 x float> %shuffle 14} 15 16define <16 x float> @shuffle_v16f32_08_08_08_08_08_08_08_08_08_08_08_08_08_08_08_08(<16 x float> %a, <16 x float> %b) { 17; ALL-LABEL: shuffle_v16f32_08_08_08_08_08_08_08_08_08_08_08_08_08_08_08_08: 18; ALL: # BB#0: 19; ALL-NEXT: vextractf32x4 $2, %zmm0, %xmm0 20; ALL-NEXT: vbroadcastss %xmm0, %zmm0 21; ALL-NEXT: retq 22 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32><i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8, i32 8> 23 ret <16 x float> %shuffle 24} 25 26define <16 x float> @shuffle_v16f32_00_10_01_11_04_14_05_15_08_18_09_19_0c_1c_0d_1d(<16 x float> %a, <16 x float> %b) { 27; ALL-LABEL: shuffle_v16f32_00_10_01_11_04_14_05_15_08_18_09_19_0c_1c_0d_1d: 28; ALL: # BB#0: 29; ALL-NEXT: vunpcklps {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13] 30; ALL-NEXT: retq 31 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32><i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 8, i32 24, i32 9, i32 25, i32 12, i32 28, i32 13, i32 29> 32 ret <16 x float> %shuffle 33} 34 35define <16 x float> @shuffle_v16f32_00_zz_01_zz_04_zz_05_zz_08_zz_09_zz_0c_zz_0d_zz(<16 x float> %a, <16 x float> %b) { 36; ALL-LABEL: shuffle_v16f32_00_zz_01_zz_04_zz_05_zz_08_zz_09_zz_0c_zz_0d_zz: 37; ALL: # BB#0: 38; ALL-NEXT: vpxord %zmm1, %zmm1, %zmm1 39; ALL-NEXT: vunpcklps {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13] 40; ALL-NEXT: retq 41 %shuffle = shufflevector <16 x float> %a, <16 x float> zeroinitializer, <16 x i32><i32 0, i32 16, i32 1, i32 16, i32 4, i32 16, i32 5, i32 16, i32 8, i32 16, i32 9, i32 16, i32 12, i32 16, i32 13, i32 16> 42 ret <16 x float> %shuffle 43} 44 45define <16 x float> @shuffle_v16f32_vunpcklps_swap(<16 x float> %a, <16 x float> %b) { 46; ALL-LABEL: shuffle_v16f32_vunpcklps_swap: 47; ALL: # BB#0: 48; ALL-NEXT: vunpcklps {{.*#+}} zmm0 = zmm1[0],zmm0[0],zmm1[1],zmm0[1],zmm1[4],zmm0[4],zmm1[5],zmm0[5],zmm1[8],zmm0[8],zmm1[9],zmm0[9],zmm1[12],zmm0[12],zmm1[13],zmm0[13] 49; ALL-NEXT: retq 50 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 16, i32 0, i32 17, i32 1, i32 20, i32 4, i32 21, i32 5, i32 24, i32 8, i32 25, i32 9, i32 28, i32 12, i32 29, i32 13> 51 ret <16 x float> %shuffle 52} 53 54define <16 x i32> @shuffle_v16i32_00_10_01_11_04_14_05_15_08_18_09_19_0c_1c_0d_1d(<16 x i32> %a, <16 x i32> %b) { 55; ALL-LABEL: shuffle_v16i32_00_10_01_11_04_14_05_15_08_18_09_19_0c_1c_0d_1d: 56; ALL: # BB#0: 57; ALL-NEXT: vpunpckldq {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13] 58; ALL-NEXT: retq 59 %shuffle = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32><i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 8, i32 24, i32 9, i32 25, i32 12, i32 28, i32 13, i32 29> 60 ret <16 x i32> %shuffle 61} 62 63define <16 x i32> @shuffle_v16i32_zz_10_zz_11_zz_14_zz_15_zz_18_zz_19_zz_1c_zz_1d(<16 x i32> %a, <16 x i32> %b) { 64; ALL-LABEL: shuffle_v16i32_zz_10_zz_11_zz_14_zz_15_zz_18_zz_19_zz_1c_zz_1d: 65; ALL: # BB#0: 66; ALL-NEXT: vpxord %zmm0, %zmm0, %zmm0 67; ALL-NEXT: vpunpckldq {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13] 68; ALL-NEXT: retq 69 %shuffle = shufflevector <16 x i32> zeroinitializer, <16 x i32> %b, <16 x i32><i32 15, i32 16, i32 13, i32 17, i32 11, i32 20, i32 9, i32 21, i32 7, i32 24, i32 5, i32 25, i32 3, i32 28, i32 1, i32 29> 70 ret <16 x i32> %shuffle 71} 72 73define <16 x float> @shuffle_v16f32_02_12_03_13_06_16_07_17_0a_1a_0b_1b_0e_1e_0f_1f(<16 x float> %a, <16 x float> %b) { 74; ALL-LABEL: shuffle_v16f32_02_12_03_13_06_16_07_17_0a_1a_0b_1b_0e_1e_0f_1f: 75; ALL: # BB#0: 76; ALL-NEXT: vunpckhps {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15] 77; ALL-NEXT: retq 78 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32><i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23, i32 10, i32 26, i32 11, i32 27, i32 14, i32 30, i32 15, i32 31> 79 ret <16 x float> %shuffle 80} 81 82define <16 x float> @shuffle_v16f32_zz_12_zz_13_zz_16_zz_17_zz_1a_zz_1b_zz_1e_zz_1f(<16 x float> %a, <16 x float> %b) { 83; ALL-LABEL: shuffle_v16f32_zz_12_zz_13_zz_16_zz_17_zz_1a_zz_1b_zz_1e_zz_1f: 84; ALL: # BB#0: 85; ALL-NEXT: vpxord %zmm0, %zmm0, %zmm0 86; ALL-NEXT: vunpckhps {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15] 87; ALL-NEXT: retq 88 %shuffle = shufflevector <16 x float> zeroinitializer, <16 x float> %b, <16 x i32><i32 0, i32 18, i32 0, i32 19, i32 4, i32 22, i32 4, i32 23, i32 6, i32 26, i32 6, i32 27, i32 8, i32 30, i32 8, i32 31> 89 ret <16 x float> %shuffle 90} 91 92define <16 x float> @shuffle_v16f32_00_00_02_02_04_04_06_06_08_08_10_10_12_12_14_14(<16 x float> %a, <16 x float> %b) { 93; ALL-LABEL: shuffle_v16f32_00_00_02_02_04_04_06_06_08_08_10_10_12_12_14_14: 94; ALL: # BB#0: 95; ALL-NEXT: vmovsldup {{.*#+}} zmm0 = zmm0[0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14] 96; ALL-NEXT: retq 97 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32><i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6, i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14> 98 ret <16 x float> %shuffle 99} 100 101define <16 x float> @shuffle_v16f32_01_01_03_03_05_05_07_07_09_09_11_11_13_13_15_15(<16 x float> %a, <16 x float> %b) { 102; ALL-LABEL: shuffle_v16f32_01_01_03_03_05_05_07_07_09_09_11_11_13_13_15_15: 103; ALL: # BB#0: 104; ALL-NEXT: vmovshdup {{.*#+}} zmm0 = zmm0[1,1,3,3,5,5,7,7,9,9,11,11,13,13,15,15] 105; ALL-NEXT: retq 106 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32><i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7, i32 9, i32 9, i32 11, i32 11, i32 13, i32 13, i32 15, i32 15> 107 ret <16 x float> %shuffle 108} 109 110define <16 x float> @shuffle_v16f32_00_01_00_01_06_07_06_07_08_09_10_11_12_13_12_13(<16 x float> %a, <16 x float> %b) { 111; ALL-LABEL: shuffle_v16f32_00_01_00_01_06_07_06_07_08_09_10_11_12_13_12_13: 112; ALL: # BB#0: 113; ALL-NEXT: vpermilpd {{.*#+}} zmm0 = zmm0[0,0,3,3,4,5,6,6] 114; ALL-NEXT: retq 115 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 1, i32 0, i32 1, i32 6, i32 7, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 12, i32 13> 116 ret <16 x float> %shuffle 117} 118 119define <16 x float> @shuffle_v16f32_00_00_02_00_04_04_06_04_08_08_10_08_12_12_14_12(<16 x float> %a, <16 x float> %b) { 120; ALL-LABEL: shuffle_v16f32_00_00_02_00_04_04_06_04_08_08_10_08_12_12_14_12: 121; ALL: # BB#0: 122; ALL-NEXT: vpermilps {{.*#+}} zmm0 = zmm0[0,0,2,0,4,4,6,4,8,8,10,8,12,12,14,12] 123; ALL-NEXT: retq 124 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 0, i32 2, i32 0, i32 4, i32 4, i32 6, i32 4, i32 8, i32 8, i32 10, i32 8, i32 12, i32 12, i32 14, i32 12> 125 ret <16 x float> %shuffle 126} 127 128define <16 x float> @shuffle_v16f32_03_uu_uu_uu_uu_04_uu_uu_uu_uu_11_uu_uu_uu_uu_12(<16 x float> %a, <16 x float> %b) { 129; ALL-LABEL: shuffle_v16f32_03_uu_uu_uu_uu_04_uu_uu_uu_uu_11_uu_uu_uu_uu_12: 130; ALL: # BB#0: 131; ALL-NEXT: vpermilps {{.*#+}} zmm0 = zmm0[3,0,3,0,7,4,7,4,11,8,11,8,15,12,15,12] 132; ALL-NEXT: retq 133 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 4, i32 undef, i32 undef, i32 undef, i32 undef, i32 11, i32 undef, i32 undef, i32 undef, i32 undef, i32 12> 134 ret <16 x float> %shuffle 135} 136 137define <16 x i32> @shuffle_v16i32_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00(<16 x i32> %a, <16 x i32> %b) { 138; ALL-LABEL: shuffle_v16i32_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00: 139; ALL: # BB#0: 140; ALL-NEXT: vpbroadcastd %xmm0, %zmm0 141; ALL-NEXT: retq 142 %shuffle = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32><i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0> 143 ret <16 x i32> %shuffle 144} 145 146define <16 x i32> @shuffle_v16i32_04_04_04_04_04_04_04_04_04_04_04_04_04_04_04_04(<16 x i32> %a, <16 x i32> %b) { 147; ALL-LABEL: shuffle_v16i32_04_04_04_04_04_04_04_04_04_04_04_04_04_04_04_04: 148; ALL: # BB#0: 149; ALL-NEXT: vextracti32x4 $1, %zmm0, %xmm0 150; ALL-NEXT: vpbroadcastd %xmm0, %zmm0 151; ALL-NEXT: retq 152 %shuffle = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32><i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4, i32 4> 153 ret <16 x i32> %shuffle 154} 155 156define <16 x i32> @shuffle_v16i32_02_12_03_13_06_16_07_17_0a_1a_0b_1b_0e_1e_0f_1f(<16 x i32> %a, <16 x i32> %b) { 157; ALL-LABEL: shuffle_v16i32_02_12_03_13_06_16_07_17_0a_1a_0b_1b_0e_1e_0f_1f: 158; ALL: # BB#0: 159; ALL-NEXT: vpunpckhdq {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15] 160; ALL-NEXT: retq 161 %shuffle = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32><i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23, i32 10, i32 26, i32 11, i32 27, i32 14, i32 30, i32 15, i32 31> 162 ret <16 x i32> %shuffle 163} 164 165define <16 x i32> @shuffle_v16i32_02_zz_03_zz_06_zz_07_zz_0a_zz_0b_zz_0e_zz_0f_zz(<16 x i32> %a, <16 x i32> %b) { 166; ALL-LABEL: shuffle_v16i32_02_zz_03_zz_06_zz_07_zz_0a_zz_0b_zz_0e_zz_0f_zz: 167; ALL: # BB#0: 168; ALL-NEXT: vpxord %zmm1, %zmm1, %zmm1 169; ALL-NEXT: vpunpckhdq {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15] 170; ALL-NEXT: retq 171 %shuffle = shufflevector <16 x i32> %a, <16 x i32> zeroinitializer, <16 x i32><i32 2, i32 30, i32 3, i32 28, i32 6, i32 26, i32 7, i32 24, i32 10, i32 22, i32 11, i32 20, i32 14, i32 18, i32 15, i32 16> 172 ret <16 x i32> %shuffle 173} 174 175define <16 x i32> @shuffle_v16i32_01_02_03_16_05_06_07_20_09_10_11_24_13_14_15_28(<16 x i32> %a, <16 x i32> %b) { 176; AVX512F-LABEL: shuffle_v16i32_01_02_03_16_05_06_07_20_09_10_11_24_13_14_15_28: 177; AVX512F: # BB#0: 178; AVX512F-NEXT: vmovdqa32 {{.*#+}} zmm2 = [1,2,3,16,5,6,7,20,9,10,11,24,13,14,15,28] 179; AVX512F-NEXT: vpermt2d %zmm1, %zmm2, %zmm0 180; AVX512F-NEXT: retq 181; 182; AVX512BW-LABEL: shuffle_v16i32_01_02_03_16_05_06_07_20_09_10_11_24_13_14_15_28: 183; AVX512BW: # BB#0: 184; AVX512BW-NEXT: vpalignr {{.*#+}} zmm0 = zmm0[4,5,6,7,8,9,10,11,12,13,14,15],zmm1[0,1,2,3],zmm0[20,21,22,23,24,25,26,27,28,29,30,31],zmm1[16,17,18,19],zmm0[36,37,38,39,40,41,42,43,44,45,46,47],zmm1[32,33,34,35],zmm0[52,53,54,55,56,57,58,59,60,61,62,63],zmm1[48,49,50,51] 185; AVX512BW-NEXT: retq 186 %shuffle = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32><i32 1, i32 2, i32 3, i32 16, i32 5, i32 6, i32 7, i32 20, i32 9, i32 10, i32 11, i32 24, i32 13, i32 14, i32 15, i32 28> 187 ret <16 x i32> %shuffle 188} 189 190define <16 x float> @shuffle_v16f32_02_05_u_u_07_u_0a_01_00_05_u_04_07_u_0a_01(<16 x float> %a) { 191; ALL-LABEL: shuffle_v16f32_02_05_u_u_07_u_0a_01_00_05_u_04_07_u_0a_01: 192; ALL: # BB#0: 193; ALL-NEXT: vmovdqa32 {{.*#+}} zmm1 = <2,5,u,u,7,u,10,1,0,5,u,4,7,u,10,1> 194; ALL-NEXT: vpermps %zmm0, %zmm1, %zmm0 195; ALL-NEXT: retq 196 %c = shufflevector <16 x float> %a, <16 x float> undef, <16 x i32> <i32 2, i32 5, i32 undef, i32 undef, i32 7, i32 undef, i32 10, i32 1, i32 0, i32 5, i32 undef, i32 4, i32 7, i32 undef, i32 10, i32 1> 197 ret <16 x float> %c 198} 199 200define <16 x i32> @shuffle_v16i32_02_05_u_u_07_u_0a_01_00_05_u_04_07_u_0a_01(<16 x i32> %a) { 201; ALL-LABEL: shuffle_v16i32_02_05_u_u_07_u_0a_01_00_05_u_04_07_u_0a_01: 202; ALL: # BB#0: 203; ALL-NEXT: vmovdqa32 {{.*#+}} zmm1 = <2,5,u,u,7,u,10,1,0,5,u,4,7,u,10,1> 204; ALL-NEXT: vpermd %zmm0, %zmm1, %zmm0 205; ALL-NEXT: retq 206 %c = shufflevector <16 x i32> %a, <16 x i32> undef, <16 x i32> <i32 2, i32 5, i32 undef, i32 undef, i32 7, i32 undef, i32 10, i32 1, i32 0, i32 5, i32 undef, i32 4, i32 7, i32 undef, i32 10, i32 1> 207 ret <16 x i32> %c 208} 209 210define <16 x i32> @shuffle_v16i32_0f_1f_0e_16_0d_1d_04_1e_0b_1b_0a_1a_09_19_08_18(<16 x i32> %a, <16 x i32> %b) { 211; ALL-LABEL: shuffle_v16i32_0f_1f_0e_16_0d_1d_04_1e_0b_1b_0a_1a_09_19_08_18: 212; ALL: # BB#0: 213; ALL-NEXT: vmovdqa32 {{.*#+}} zmm2 = [15,31,14,22,13,29,4,28,11,27,10,26,9,25,8,24] 214; ALL-NEXT: vpermt2d %zmm1, %zmm2, %zmm0 215; ALL-NEXT: retq 216 %c = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24> 217 ret <16 x i32> %c 218} 219 220define <16 x float> @shuffle_v16f32_0f_1f_0e_16_0d_1d_04_1e_0b_1b_0a_1a_09_19_08_18(<16 x float> %a, <16 x float> %b) { 221; ALL-LABEL: shuffle_v16f32_0f_1f_0e_16_0d_1d_04_1e_0b_1b_0a_1a_09_19_08_18: 222; ALL: # BB#0: 223; ALL-NEXT: vmovdqa32 {{.*#+}} zmm2 = [15,31,14,22,13,29,4,28,11,27,10,26,9,25,8,24] 224; ALL-NEXT: vpermt2ps %zmm1, %zmm2, %zmm0 225; ALL-NEXT: retq 226 %c = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24> 227 ret <16 x float> %c 228} 229 230define <16 x float> @shuffle_v16f32_load_0f_1f_0e_16_0d_1d_04_1e_0b_1b_0a_1a_09_19_08_18(<16 x float> %a, <16 x float>* %b) { 231; ALL-LABEL: shuffle_v16f32_load_0f_1f_0e_16_0d_1d_04_1e_0b_1b_0a_1a_09_19_08_18: 232; ALL: # BB#0: 233; ALL-NEXT: vmovdqa32 {{.*#+}} zmm1 = [15,31,14,22,13,29,4,28,11,27,10,26,9,25,8,24] 234; ALL-NEXT: vpermt2ps (%rdi), %zmm1, %zmm0 235; ALL-NEXT: retq 236 %c = load <16 x float>, <16 x float>* %b 237 %d = shufflevector <16 x float> %a, <16 x float> %c, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24> 238 ret <16 x float> %d 239} 240 241define <16 x i32> @shuffle_v16i32_load_0f_1f_0e_16_0d_1d_04_1e_0b_1b_0a_1a_09_19_08_18(<16 x i32> %a, <16 x i32>* %b) { 242; ALL-LABEL: shuffle_v16i32_load_0f_1f_0e_16_0d_1d_04_1e_0b_1b_0a_1a_09_19_08_18: 243; ALL: # BB#0: 244; ALL-NEXT: vmovdqa32 {{.*#+}} zmm1 = [15,31,14,22,13,29,4,28,11,27,10,26,9,25,8,24] 245; ALL-NEXT: vpermt2d (%rdi), %zmm1, %zmm0 246; ALL-NEXT: retq 247 %c = load <16 x i32>, <16 x i32>* %b 248 %d = shufflevector <16 x i32> %a, <16 x i32> %c, <16 x i32> <i32 15, i32 31, i32 14, i32 22, i32 13, i32 29, i32 4, i32 28, i32 11, i32 27, i32 10, i32 26, i32 9, i32 25, i32 8, i32 24> 249 ret <16 x i32> %d 250} 251 252define <16 x i32> @shuffle_v16i32_0_1_2_19_u_u_u_u_u_u_u_u_u_u_u_u(<16 x i32> %a, <16 x i32> %b) { 253; ALL-LABEL: shuffle_v16i32_0_1_2_19_u_u_u_u_u_u_u_u_u_u_u_u: 254; ALL: # BB#0: 255; ALL-NEXT: vmovdqa32 {{.*#+}} zmm2 = <0,1,2,19,u,u,u,u,u,u,u,u,u,u,u,u> 256; ALL-NEXT: vpermt2d %zmm1, %zmm2, %zmm0 257; ALL-NEXT: retq 258 %c = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 0, i32 1, i32 2, i32 19, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef> 259 ret <16 x i32> %c 260} 261 262define <8 x float> @shuffle_v16f32_extract_256(float* %RET, float* %a) { 263; ALL-LABEL: shuffle_v16f32_extract_256: 264; ALL: # BB#0: 265; ALL-NEXT: vmovups (%rsi), %zmm0 266; ALL-NEXT: vextractf64x4 $1, %zmm0, %ymm0 267; ALL-NEXT: retq 268 %ptr_a = bitcast float* %a to <16 x float>* 269 %v_a = load <16 x float>, <16 x float>* %ptr_a, align 4 270 %v2 = shufflevector <16 x float> %v_a, <16 x float> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15> 271 ret <8 x float> %v2 272} 273 274define <16 x i32> @shuffle_v16i16_1_0_0_0_5_4_4_4_9_8_8_8_13_12_12_12(<16 x i32> %a, <16 x i32> %b) { 275; ALL-LABEL: shuffle_v16i16_1_0_0_0_5_4_4_4_9_8_8_8_13_12_12_12: 276; ALL: # BB#0: 277; ALL-NEXT: vpshufd {{.*#+}} zmm0 = zmm0[1,0,0,0,5,4,4,4,9,8,8,8,13,12,12,12] 278; ALL-NEXT: retq 279 %c = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 1, i32 0, i32 0, i32 0, i32 5, i32 4, i32 4, i32 4, i32 9, i32 8, i32 8, i32 8, i32 13, i32 12, i32 12, i32 12> 280 ret <16 x i32> %c 281} 282 283define <16 x i32> @shuffle_v16i16_3_3_0_0_7_7_4_4_11_11_8_8_15_15_12_12(<16 x i32> %a, <16 x i32> %b) { 284; ALL-LABEL: shuffle_v16i16_3_3_0_0_7_7_4_4_11_11_8_8_15_15_12_12: 285; ALL: # BB#0: 286; ALL-NEXT: vpshufd {{.*#+}} zmm0 = zmm0[2,3,0,1,6,7,4,5,10,11,8,9,14,15,12,13] 287; ALL-NEXT: retq 288 %c = shufflevector <16 x i32> %a, <16 x i32> %b, <16 x i32> <i32 2, i32 3, i32 0, i32 1, i32 6, i32 7, i32 4, i32 5, i32 10, i32 11, i32 8, i32 9, i32 14, i32 15, i32 12, i32 13> 289 ret <16 x i32> %c 290} 291 292define <16 x float> @shuffle_v16f32_00_01_10_10_04_05_14_14_08_09_18_18_0c_0d_1c_1c(<16 x float> %a, <16 x float> %b) { 293; ALL-LABEL: shuffle_v16f32_00_01_10_10_04_05_14_14_08_09_18_18_0c_0d_1c_1c: 294; ALL: # BB#0: 295; ALL-NEXT: vshufps {{.*#+}} zmm0 = zmm0[0,1],zmm1[0,0],zmm0[4,5],zmm1[4,4],zmm0[8,9],zmm1[8,8],zmm0[12,13],zmm1[12,12] 296; ALL-NEXT: retq 297 %shuffle = shufflevector <16 x float> %a, <16 x float> %b, <16 x i32> <i32 0, i32 1, i32 16, i32 16, i32 4, i32 5, i32 20, i32 20, i32 8, i32 9, i32 24, i32 24, i32 12, i32 13, i32 28, i32 28> 298 ret <16 x float> %shuffle 299} 300