1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py 2; RUN: opt < %s -mtriple=x86_64-unknown -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=SSE 3; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=slm -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=SSE 4; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7-avx -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX,AVX1 5; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=core-avx2 -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX,AVX2 6; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=skx -mattr=+prefer-256-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX,AVX256 7; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=skx -mattr=-prefer-256-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX512 8; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=znver4 -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX512VBMI2 9; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=knl -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX512 10 11@a64 = common global [8 x i64] zeroinitializer, align 64 12@b64 = common global [8 x i64] zeroinitializer, align 64 13@c64 = common global [8 x i64] zeroinitializer, align 64 14@d64 = common global [8 x i64] zeroinitializer, align 64 15@a32 = common global [16 x i32] zeroinitializer, align 64 16@b32 = common global [16 x i32] zeroinitializer, align 64 17@c32 = common global [16 x i32] zeroinitializer, align 64 18@d32 = common global [16 x i32] zeroinitializer, align 64 19@a16 = common global [32 x i16] zeroinitializer, align 64 20@b16 = common global [32 x i16] zeroinitializer, align 64 21@c16 = common global [32 x i16] zeroinitializer, align 64 22@d16 = common global [32 x i16] zeroinitializer, align 64 23@a8 = common global [64 x i8] zeroinitializer, align 64 24@b8 = common global [64 x i8] zeroinitializer, align 64 25@c8 = common global [64 x i8] zeroinitializer, align 64 26@d8 = common global [64 x i8] zeroinitializer, align 64 27 28declare i64 @llvm.fshl.i64(i64, i64, i64) 29declare i32 @llvm.fshl.i32(i32, i32, i32) 30declare i16 @llvm.fshl.i16(i16, i16, i16) 31declare i8 @llvm.fshl.i8 (i8 , i8 , i8 ) 32 33define void @fshl_v8i64() { 34; SSE-LABEL: @fshl_v8i64( 35; SSE-NEXT: [[A0:%.*]] = load i64, ptr @a64, align 8 36; SSE-NEXT: [[A1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 8 37; SSE-NEXT: [[A2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8 38; SSE-NEXT: [[A3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 8 39; SSE-NEXT: [[A4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8 40; SSE-NEXT: [[A5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 8 41; SSE-NEXT: [[A6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8 42; SSE-NEXT: [[A7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 8 43; SSE-NEXT: [[B0:%.*]] = load i64, ptr @b64, align 8 44; SSE-NEXT: [[B1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 8 45; SSE-NEXT: [[B2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8 46; SSE-NEXT: [[B3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 8 47; SSE-NEXT: [[B4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8 48; SSE-NEXT: [[B5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 8 49; SSE-NEXT: [[B6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8 50; SSE-NEXT: [[B7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 8 51; SSE-NEXT: [[C0:%.*]] = load i64, ptr @c64, align 8 52; SSE-NEXT: [[C1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 1), align 8 53; SSE-NEXT: [[C2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8 54; SSE-NEXT: [[C3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 3), align 8 55; SSE-NEXT: [[C4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8 56; SSE-NEXT: [[C5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 5), align 8 57; SSE-NEXT: [[C6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8 58; SSE-NEXT: [[C7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 7), align 8 59; SSE-NEXT: [[R0:%.*]] = call i64 @llvm.fshl.i64(i64 [[A0]], i64 [[B0]], i64 [[C0]]) 60; SSE-NEXT: [[R1:%.*]] = call i64 @llvm.fshl.i64(i64 [[A1]], i64 [[B1]], i64 [[C1]]) 61; SSE-NEXT: [[R2:%.*]] = call i64 @llvm.fshl.i64(i64 [[A2]], i64 [[B2]], i64 [[C2]]) 62; SSE-NEXT: [[R3:%.*]] = call i64 @llvm.fshl.i64(i64 [[A3]], i64 [[B3]], i64 [[C3]]) 63; SSE-NEXT: [[R4:%.*]] = call i64 @llvm.fshl.i64(i64 [[A4]], i64 [[B4]], i64 [[C4]]) 64; SSE-NEXT: [[R5:%.*]] = call i64 @llvm.fshl.i64(i64 [[A5]], i64 [[B5]], i64 [[C5]]) 65; SSE-NEXT: [[R6:%.*]] = call i64 @llvm.fshl.i64(i64 [[A6]], i64 [[B6]], i64 [[C6]]) 66; SSE-NEXT: [[R7:%.*]] = call i64 @llvm.fshl.i64(i64 [[A7]], i64 [[B7]], i64 [[C7]]) 67; SSE-NEXT: store i64 [[R0]], ptr @d64, align 8 68; SSE-NEXT: store i64 [[R1]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 1), align 8 69; SSE-NEXT: store i64 [[R2]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 2), align 8 70; SSE-NEXT: store i64 [[R3]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 3), align 8 71; SSE-NEXT: store i64 [[R4]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 8 72; SSE-NEXT: store i64 [[R5]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 5), align 8 73; SSE-NEXT: store i64 [[R6]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 6), align 8 74; SSE-NEXT: store i64 [[R7]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 7), align 8 75; SSE-NEXT: ret void 76; 77; AVX1-LABEL: @fshl_v8i64( 78; AVX1-NEXT: [[TMP1:%.*]] = load <2 x i64>, ptr @a64, align 8 79; AVX1-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @b64, align 8 80; AVX1-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr @c64, align 8 81; AVX1-NEXT: [[TMP4:%.*]] = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> [[TMP1]], <2 x i64> [[TMP2]], <2 x i64> [[TMP3]]) 82; AVX1-NEXT: store <2 x i64> [[TMP4]], ptr @d64, align 8 83; AVX1-NEXT: [[TMP5:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8 84; AVX1-NEXT: [[TMP6:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8 85; AVX1-NEXT: [[TMP7:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8 86; AVX1-NEXT: [[TMP8:%.*]] = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> [[TMP5]], <2 x i64> [[TMP6]], <2 x i64> [[TMP7]]) 87; AVX1-NEXT: store <2 x i64> [[TMP8]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 2), align 8 88; AVX1-NEXT: [[TMP9:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8 89; AVX1-NEXT: [[TMP10:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8 90; AVX1-NEXT: [[TMP11:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8 91; AVX1-NEXT: [[TMP12:%.*]] = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> [[TMP9]], <2 x i64> [[TMP10]], <2 x i64> [[TMP11]]) 92; AVX1-NEXT: store <2 x i64> [[TMP12]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 8 93; AVX1-NEXT: [[TMP13:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8 94; AVX1-NEXT: [[TMP14:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8 95; AVX1-NEXT: [[TMP15:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8 96; AVX1-NEXT: [[TMP16:%.*]] = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> [[TMP13]], <2 x i64> [[TMP14]], <2 x i64> [[TMP15]]) 97; AVX1-NEXT: store <2 x i64> [[TMP16]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 6), align 8 98; AVX1-NEXT: ret void 99; 100; AVX2-LABEL: @fshl_v8i64( 101; AVX2-NEXT: [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 8 102; AVX2-NEXT: [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 8 103; AVX2-NEXT: [[TMP3:%.*]] = load <4 x i64>, ptr @c64, align 8 104; AVX2-NEXT: [[TMP4:%.*]] = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> [[TMP1]], <4 x i64> [[TMP2]], <4 x i64> [[TMP3]]) 105; AVX2-NEXT: store <4 x i64> [[TMP4]], ptr @d64, align 8 106; AVX2-NEXT: [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8 107; AVX2-NEXT: [[TMP6:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8 108; AVX2-NEXT: [[TMP7:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8 109; AVX2-NEXT: [[TMP8:%.*]] = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> [[TMP5]], <4 x i64> [[TMP6]], <4 x i64> [[TMP7]]) 110; AVX2-NEXT: store <4 x i64> [[TMP8]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 8 111; AVX2-NEXT: ret void 112; 113; AVX256-LABEL: @fshl_v8i64( 114; AVX256-NEXT: [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 8 115; AVX256-NEXT: [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 8 116; AVX256-NEXT: [[TMP3:%.*]] = load <4 x i64>, ptr @c64, align 8 117; AVX256-NEXT: [[TMP4:%.*]] = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> [[TMP1]], <4 x i64> [[TMP2]], <4 x i64> [[TMP3]]) 118; AVX256-NEXT: store <4 x i64> [[TMP4]], ptr @d64, align 8 119; AVX256-NEXT: [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8 120; AVX256-NEXT: [[TMP6:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8 121; AVX256-NEXT: [[TMP7:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8 122; AVX256-NEXT: [[TMP8:%.*]] = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> [[TMP5]], <4 x i64> [[TMP6]], <4 x i64> [[TMP7]]) 123; AVX256-NEXT: store <4 x i64> [[TMP8]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 8 124; AVX256-NEXT: ret void 125; 126; AVX512-LABEL: @fshl_v8i64( 127; AVX512-NEXT: [[TMP1:%.*]] = load <8 x i64>, ptr @a64, align 8 128; AVX512-NEXT: [[TMP2:%.*]] = load <8 x i64>, ptr @b64, align 8 129; AVX512-NEXT: [[TMP3:%.*]] = load <8 x i64>, ptr @c64, align 8 130; AVX512-NEXT: [[TMP4:%.*]] = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> [[TMP1]], <8 x i64> [[TMP2]], <8 x i64> [[TMP3]]) 131; AVX512-NEXT: store <8 x i64> [[TMP4]], ptr @d64, align 8 132; AVX512-NEXT: ret void 133; 134; AVX512VBMI2-LABEL: @fshl_v8i64( 135; AVX512VBMI2-NEXT: [[TMP1:%.*]] = load <8 x i64>, ptr @a64, align 8 136; AVX512VBMI2-NEXT: [[TMP2:%.*]] = load <8 x i64>, ptr @b64, align 8 137; AVX512VBMI2-NEXT: [[TMP3:%.*]] = load <8 x i64>, ptr @c64, align 8 138; AVX512VBMI2-NEXT: [[TMP4:%.*]] = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> [[TMP1]], <8 x i64> [[TMP2]], <8 x i64> [[TMP3]]) 139; AVX512VBMI2-NEXT: store <8 x i64> [[TMP4]], ptr @d64, align 8 140; AVX512VBMI2-NEXT: ret void 141; 142 %a0 = load i64, ptr @a64, align 8 143 %a1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 8 144 %a2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8 145 %a3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 8 146 %a4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8 147 %a5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 8 148 %a6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8 149 %a7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 8 150 %b0 = load i64, ptr @b64, align 8 151 %b1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 8 152 %b2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8 153 %b3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 8 154 %b4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8 155 %b5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 8 156 %b6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8 157 %b7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 8 158 %c0 = load i64, ptr @c64, align 8 159 %c1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 1), align 8 160 %c2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8 161 %c3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 3), align 8 162 %c4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8 163 %c5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 5), align 8 164 %c6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8 165 %c7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 7), align 8 166 %r0 = call i64 @llvm.fshl.i64(i64 %a0, i64 %b0, i64 %c0) 167 %r1 = call i64 @llvm.fshl.i64(i64 %a1, i64 %b1, i64 %c1) 168 %r2 = call i64 @llvm.fshl.i64(i64 %a2, i64 %b2, i64 %c2) 169 %r3 = call i64 @llvm.fshl.i64(i64 %a3, i64 %b3, i64 %c3) 170 %r4 = call i64 @llvm.fshl.i64(i64 %a4, i64 %b4, i64 %c4) 171 %r5 = call i64 @llvm.fshl.i64(i64 %a5, i64 %b5, i64 %c5) 172 %r6 = call i64 @llvm.fshl.i64(i64 %a6, i64 %b6, i64 %c6) 173 %r7 = call i64 @llvm.fshl.i64(i64 %a7, i64 %b7, i64 %c7) 174 store i64 %r0, ptr @d64, align 8 175 store i64 %r1, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 1), align 8 176 store i64 %r2, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 2), align 8 177 store i64 %r3, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 3), align 8 178 store i64 %r4, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 8 179 store i64 %r5, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 5), align 8 180 store i64 %r6, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 6), align 8 181 store i64 %r7, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 7), align 8 182 ret void 183} 184 185define void @fshl_v16i32() { 186; SSE-LABEL: @fshl_v16i32( 187; SSE-NEXT: [[A0:%.*]] = load i32, ptr @a32, align 4 188; SSE-NEXT: [[A1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1), align 4 189; SSE-NEXT: [[A2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2), align 4 190; SSE-NEXT: [[A3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3), align 4 191; SSE-NEXT: [[A4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4 192; SSE-NEXT: [[A5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5), align 4 193; SSE-NEXT: [[A6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6), align 4 194; SSE-NEXT: [[A7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7), align 4 195; SSE-NEXT: [[A8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4 196; SSE-NEXT: [[A9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9), align 4 197; SSE-NEXT: [[A10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 4 198; SSE-NEXT: [[A11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 4 199; SSE-NEXT: [[A12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4 200; SSE-NEXT: [[A13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 4 201; SSE-NEXT: [[A14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 4 202; SSE-NEXT: [[A15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 4 203; SSE-NEXT: [[B0:%.*]] = load i32, ptr @b32, align 4 204; SSE-NEXT: [[B1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1), align 4 205; SSE-NEXT: [[B2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2), align 4 206; SSE-NEXT: [[B3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3), align 4 207; SSE-NEXT: [[B4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4 208; SSE-NEXT: [[B5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5), align 4 209; SSE-NEXT: [[B6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6), align 4 210; SSE-NEXT: [[B7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7), align 4 211; SSE-NEXT: [[B8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4 212; SSE-NEXT: [[B9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9), align 4 213; SSE-NEXT: [[B10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 4 214; SSE-NEXT: [[B11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 4 215; SSE-NEXT: [[B12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4 216; SSE-NEXT: [[B13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 4 217; SSE-NEXT: [[B14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 4 218; SSE-NEXT: [[B15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 4 219; SSE-NEXT: [[C0:%.*]] = load i32, ptr @c32, align 4 220; SSE-NEXT: [[C1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 1), align 4 221; SSE-NEXT: [[C2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 2), align 4 222; SSE-NEXT: [[C3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 3), align 4 223; SSE-NEXT: [[C4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4), align 4 224; SSE-NEXT: [[C5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 5), align 4 225; SSE-NEXT: [[C6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 6), align 4 226; SSE-NEXT: [[C7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 7), align 4 227; SSE-NEXT: [[C8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4 228; SSE-NEXT: [[C9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 9), align 4 229; SSE-NEXT: [[C10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 10), align 4 230; SSE-NEXT: [[C11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 11), align 4 231; SSE-NEXT: [[C12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4 232; SSE-NEXT: [[C13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 13), align 4 233; SSE-NEXT: [[C14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 14), align 4 234; SSE-NEXT: [[C15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 15), align 4 235; SSE-NEXT: [[R0:%.*]] = call i32 @llvm.fshl.i32(i32 [[A0]], i32 [[B0]], i32 [[C0]]) 236; SSE-NEXT: [[R1:%.*]] = call i32 @llvm.fshl.i32(i32 [[A1]], i32 [[B1]], i32 [[C1]]) 237; SSE-NEXT: [[R2:%.*]] = call i32 @llvm.fshl.i32(i32 [[A2]], i32 [[B2]], i32 [[C2]]) 238; SSE-NEXT: [[R3:%.*]] = call i32 @llvm.fshl.i32(i32 [[A3]], i32 [[B3]], i32 [[C3]]) 239; SSE-NEXT: [[R4:%.*]] = call i32 @llvm.fshl.i32(i32 [[A4]], i32 [[B4]], i32 [[C4]]) 240; SSE-NEXT: [[R5:%.*]] = call i32 @llvm.fshl.i32(i32 [[A5]], i32 [[B5]], i32 [[C5]]) 241; SSE-NEXT: [[R6:%.*]] = call i32 @llvm.fshl.i32(i32 [[A6]], i32 [[B6]], i32 [[C6]]) 242; SSE-NEXT: [[R7:%.*]] = call i32 @llvm.fshl.i32(i32 [[A7]], i32 [[B7]], i32 [[C7]]) 243; SSE-NEXT: [[R8:%.*]] = call i32 @llvm.fshl.i32(i32 [[A8]], i32 [[B8]], i32 [[C8]]) 244; SSE-NEXT: [[R9:%.*]] = call i32 @llvm.fshl.i32(i32 [[A9]], i32 [[B9]], i32 [[C9]]) 245; SSE-NEXT: [[R10:%.*]] = call i32 @llvm.fshl.i32(i32 [[A10]], i32 [[B10]], i32 [[C10]]) 246; SSE-NEXT: [[R11:%.*]] = call i32 @llvm.fshl.i32(i32 [[A11]], i32 [[B11]], i32 [[C11]]) 247; SSE-NEXT: [[R12:%.*]] = call i32 @llvm.fshl.i32(i32 [[A12]], i32 [[B12]], i32 [[C12]]) 248; SSE-NEXT: [[R13:%.*]] = call i32 @llvm.fshl.i32(i32 [[A13]], i32 [[B13]], i32 [[C13]]) 249; SSE-NEXT: [[R14:%.*]] = call i32 @llvm.fshl.i32(i32 [[A14]], i32 [[B14]], i32 [[C14]]) 250; SSE-NEXT: [[R15:%.*]] = call i32 @llvm.fshl.i32(i32 [[A15]], i32 [[B15]], i32 [[C15]]) 251; SSE-NEXT: store i32 [[R0]], ptr @d32, align 4 252; SSE-NEXT: store i32 [[R1]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 1), align 4 253; SSE-NEXT: store i32 [[R2]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 2), align 4 254; SSE-NEXT: store i32 [[R3]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 3), align 4 255; SSE-NEXT: store i32 [[R4]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 4), align 4 256; SSE-NEXT: store i32 [[R5]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 5), align 4 257; SSE-NEXT: store i32 [[R6]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 6), align 4 258; SSE-NEXT: store i32 [[R7]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 7), align 4 259; SSE-NEXT: store i32 [[R8]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 8), align 4 260; SSE-NEXT: store i32 [[R9]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 9), align 4 261; SSE-NEXT: store i32 [[R10]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 10), align 4 262; SSE-NEXT: store i32 [[R11]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 11), align 4 263; SSE-NEXT: store i32 [[R12]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 12), align 4 264; SSE-NEXT: store i32 [[R13]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 13), align 4 265; SSE-NEXT: store i32 [[R14]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 14), align 4 266; SSE-NEXT: store i32 [[R15]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 15), align 4 267; SSE-NEXT: ret void 268; 269; AVX-LABEL: @fshl_v16i32( 270; AVX-NEXT: [[TMP1:%.*]] = load <8 x i32>, ptr @a32, align 4 271; AVX-NEXT: [[TMP2:%.*]] = load <8 x i32>, ptr @b32, align 4 272; AVX-NEXT: [[TMP3:%.*]] = load <8 x i32>, ptr @c32, align 4 273; AVX-NEXT: [[TMP4:%.*]] = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> [[TMP1]], <8 x i32> [[TMP2]], <8 x i32> [[TMP3]]) 274; AVX-NEXT: store <8 x i32> [[TMP4]], ptr @d32, align 4 275; AVX-NEXT: [[TMP5:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4 276; AVX-NEXT: [[TMP6:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4 277; AVX-NEXT: [[TMP7:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4 278; AVX-NEXT: [[TMP8:%.*]] = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> [[TMP5]], <8 x i32> [[TMP6]], <8 x i32> [[TMP7]]) 279; AVX-NEXT: store <8 x i32> [[TMP8]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 8), align 4 280; AVX-NEXT: ret void 281; 282; AVX512-LABEL: @fshl_v16i32( 283; AVX512-NEXT: [[TMP1:%.*]] = load <16 x i32>, ptr @a32, align 4 284; AVX512-NEXT: [[TMP2:%.*]] = load <16 x i32>, ptr @b32, align 4 285; AVX512-NEXT: [[TMP3:%.*]] = load <16 x i32>, ptr @c32, align 4 286; AVX512-NEXT: [[TMP4:%.*]] = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> [[TMP1]], <16 x i32> [[TMP2]], <16 x i32> [[TMP3]]) 287; AVX512-NEXT: store <16 x i32> [[TMP4]], ptr @d32, align 4 288; AVX512-NEXT: ret void 289; 290; AVX512VBMI2-LABEL: @fshl_v16i32( 291; AVX512VBMI2-NEXT: [[TMP1:%.*]] = load <16 x i32>, ptr @a32, align 4 292; AVX512VBMI2-NEXT: [[TMP2:%.*]] = load <16 x i32>, ptr @b32, align 4 293; AVX512VBMI2-NEXT: [[TMP3:%.*]] = load <16 x i32>, ptr @c32, align 4 294; AVX512VBMI2-NEXT: [[TMP4:%.*]] = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> [[TMP1]], <16 x i32> [[TMP2]], <16 x i32> [[TMP3]]) 295; AVX512VBMI2-NEXT: store <16 x i32> [[TMP4]], ptr @d32, align 4 296; AVX512VBMI2-NEXT: ret void 297; 298 %a0 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 0 ), align 4 299 %a1 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1 ), align 4 300 %a2 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2 ), align 4 301 %a3 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3 ), align 4 302 %a4 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4 ), align 4 303 %a5 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5 ), align 4 304 %a6 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6 ), align 4 305 %a7 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7 ), align 4 306 %a8 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8 ), align 4 307 %a9 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9 ), align 4 308 %a10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 4 309 %a11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 4 310 %a12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4 311 %a13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 4 312 %a14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 4 313 %a15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 4 314 %b0 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 0 ), align 4 315 %b1 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1 ), align 4 316 %b2 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2 ), align 4 317 %b3 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3 ), align 4 318 %b4 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4 ), align 4 319 %b5 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5 ), align 4 320 %b6 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6 ), align 4 321 %b7 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7 ), align 4 322 %b8 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8 ), align 4 323 %b9 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9 ), align 4 324 %b10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 4 325 %b11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 4 326 %b12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4 327 %b13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 4 328 %b14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 4 329 %b15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 4 330 %c0 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 0 ), align 4 331 %c1 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 1 ), align 4 332 %c2 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 2 ), align 4 333 %c3 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 3 ), align 4 334 %c4 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4 ), align 4 335 %c5 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 5 ), align 4 336 %c6 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 6 ), align 4 337 %c7 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 7 ), align 4 338 %c8 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8 ), align 4 339 %c9 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 9 ), align 4 340 %c10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 10), align 4 341 %c11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 11), align 4 342 %c12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4 343 %c13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 13), align 4 344 %c14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 14), align 4 345 %c15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 15), align 4 346 %r0 = call i32 @llvm.fshl.i32(i32 %a0 , i32 %b0 , i32 %c0 ) 347 %r1 = call i32 @llvm.fshl.i32(i32 %a1 , i32 %b1 , i32 %c1 ) 348 %r2 = call i32 @llvm.fshl.i32(i32 %a2 , i32 %b2 , i32 %c2 ) 349 %r3 = call i32 @llvm.fshl.i32(i32 %a3 , i32 %b3 , i32 %c3 ) 350 %r4 = call i32 @llvm.fshl.i32(i32 %a4 , i32 %b4 , i32 %c4 ) 351 %r5 = call i32 @llvm.fshl.i32(i32 %a5 , i32 %b5 , i32 %c5 ) 352 %r6 = call i32 @llvm.fshl.i32(i32 %a6 , i32 %b6 , i32 %c6 ) 353 %r7 = call i32 @llvm.fshl.i32(i32 %a7 , i32 %b7 , i32 %c7 ) 354 %r8 = call i32 @llvm.fshl.i32(i32 %a8 , i32 %b8 , i32 %c8 ) 355 %r9 = call i32 @llvm.fshl.i32(i32 %a9 , i32 %b9 , i32 %c9 ) 356 %r10 = call i32 @llvm.fshl.i32(i32 %a10, i32 %b10, i32 %c10) 357 %r11 = call i32 @llvm.fshl.i32(i32 %a11, i32 %b11, i32 %c11) 358 %r12 = call i32 @llvm.fshl.i32(i32 %a12, i32 %b12, i32 %c12) 359 %r13 = call i32 @llvm.fshl.i32(i32 %a13, i32 %b13, i32 %c13) 360 %r14 = call i32 @llvm.fshl.i32(i32 %a14, i32 %b14, i32 %c14) 361 %r15 = call i32 @llvm.fshl.i32(i32 %a15, i32 %b15, i32 %c15) 362 store i32 %r0 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 0 ), align 4 363 store i32 %r1 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 1 ), align 4 364 store i32 %r2 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 2 ), align 4 365 store i32 %r3 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 3 ), align 4 366 store i32 %r4 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 4 ), align 4 367 store i32 %r5 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 5 ), align 4 368 store i32 %r6 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 6 ), align 4 369 store i32 %r7 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 7 ), align 4 370 store i32 %r8 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 8 ), align 4 371 store i32 %r9 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 9 ), align 4 372 store i32 %r10, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 10), align 4 373 store i32 %r11, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 11), align 4 374 store i32 %r12, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 12), align 4 375 store i32 %r13, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 13), align 4 376 store i32 %r14, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 14), align 4 377 store i32 %r15, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 15), align 4 378 ret void 379} 380 381define void @fshl_v32i16() { 382; SSE-LABEL: @fshl_v32i16( 383; SSE-NEXT: [[TMP1:%.*]] = load <8 x i16>, ptr @a16, align 2 384; SSE-NEXT: [[TMP2:%.*]] = load <8 x i16>, ptr @b16, align 2 385; SSE-NEXT: [[TMP3:%.*]] = load <8 x i16>, ptr @c16, align 2 386; SSE-NEXT: [[TMP4:%.*]] = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> [[TMP1]], <8 x i16> [[TMP2]], <8 x i16> [[TMP3]]) 387; SSE-NEXT: store <8 x i16> [[TMP4]], ptr @d16, align 2 388; SSE-NEXT: [[TMP5:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8), align 2 389; SSE-NEXT: [[TMP6:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8), align 2 390; SSE-NEXT: [[TMP7:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8), align 2 391; SSE-NEXT: [[TMP8:%.*]] = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> [[TMP5]], <8 x i16> [[TMP6]], <8 x i16> [[TMP7]]) 392; SSE-NEXT: store <8 x i16> [[TMP8]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 8), align 2 393; SSE-NEXT: [[TMP9:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2 394; SSE-NEXT: [[TMP10:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2 395; SSE-NEXT: [[TMP11:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2 396; SSE-NEXT: [[TMP12:%.*]] = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> [[TMP9]], <8 x i16> [[TMP10]], <8 x i16> [[TMP11]]) 397; SSE-NEXT: store <8 x i16> [[TMP12]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 16), align 2 398; SSE-NEXT: [[TMP13:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2 399; SSE-NEXT: [[TMP14:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2 400; SSE-NEXT: [[TMP15:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2 401; SSE-NEXT: [[TMP16:%.*]] = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> [[TMP13]], <8 x i16> [[TMP14]], <8 x i16> [[TMP15]]) 402; SSE-NEXT: store <8 x i16> [[TMP16]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 24), align 2 403; SSE-NEXT: ret void 404; 405; AVX-LABEL: @fshl_v32i16( 406; AVX-NEXT: [[TMP1:%.*]] = load <16 x i16>, ptr @a16, align 2 407; AVX-NEXT: [[TMP2:%.*]] = load <16 x i16>, ptr @b16, align 2 408; AVX-NEXT: [[TMP3:%.*]] = load <16 x i16>, ptr @c16, align 2 409; AVX-NEXT: [[TMP4:%.*]] = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> [[TMP1]], <16 x i16> [[TMP2]], <16 x i16> [[TMP3]]) 410; AVX-NEXT: store <16 x i16> [[TMP4]], ptr @d16, align 2 411; AVX-NEXT: [[TMP5:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2 412; AVX-NEXT: [[TMP6:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2 413; AVX-NEXT: [[TMP7:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2 414; AVX-NEXT: [[TMP8:%.*]] = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> [[TMP5]], <16 x i16> [[TMP6]], <16 x i16> [[TMP7]]) 415; AVX-NEXT: store <16 x i16> [[TMP8]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 16), align 2 416; AVX-NEXT: ret void 417; 418; AVX512-LABEL: @fshl_v32i16( 419; AVX512-NEXT: [[TMP1:%.*]] = load <32 x i16>, ptr @a16, align 2 420; AVX512-NEXT: [[TMP2:%.*]] = load <32 x i16>, ptr @b16, align 2 421; AVX512-NEXT: [[TMP3:%.*]] = load <32 x i16>, ptr @c16, align 2 422; AVX512-NEXT: [[TMP4:%.*]] = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> [[TMP1]], <32 x i16> [[TMP2]], <32 x i16> [[TMP3]]) 423; AVX512-NEXT: store <32 x i16> [[TMP4]], ptr @d16, align 2 424; AVX512-NEXT: ret void 425; 426; AVX512VBMI2-LABEL: @fshl_v32i16( 427; AVX512VBMI2-NEXT: [[TMP1:%.*]] = load <32 x i16>, ptr @a16, align 2 428; AVX512VBMI2-NEXT: [[TMP2:%.*]] = load <32 x i16>, ptr @b16, align 2 429; AVX512VBMI2-NEXT: [[TMP3:%.*]] = load <32 x i16>, ptr @c16, align 2 430; AVX512VBMI2-NEXT: [[TMP4:%.*]] = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> [[TMP1]], <32 x i16> [[TMP2]], <32 x i16> [[TMP3]]) 431; AVX512VBMI2-NEXT: store <32 x i16> [[TMP4]], ptr @d16, align 2 432; AVX512VBMI2-NEXT: ret void 433; 434 %a0 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 0 ), align 2 435 %a1 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 1 ), align 2 436 %a2 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 2 ), align 2 437 %a3 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 3 ), align 2 438 %a4 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 4 ), align 2 439 %a5 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 5 ), align 2 440 %a6 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 6 ), align 2 441 %a7 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 7 ), align 2 442 %a8 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8 ), align 2 443 %a9 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 9 ), align 2 444 %a10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 10), align 2 445 %a11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 11), align 2 446 %a12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 12), align 2 447 %a13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 13), align 2 448 %a14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 14), align 2 449 %a15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 15), align 2 450 %a16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2 451 %a17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 17), align 2 452 %a18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 18), align 2 453 %a19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 19), align 2 454 %a20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 20), align 2 455 %a21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 21), align 2 456 %a22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 22), align 2 457 %a23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 23), align 2 458 %a24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2 459 %a25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 25), align 2 460 %a26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 26), align 2 461 %a27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 27), align 2 462 %a28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 28), align 2 463 %a29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 29), align 2 464 %a30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 30), align 2 465 %a31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 31), align 2 466 %b0 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 0 ), align 2 467 %b1 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 1 ), align 2 468 %b2 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 2 ), align 2 469 %b3 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 3 ), align 2 470 %b4 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 4 ), align 2 471 %b5 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 5 ), align 2 472 %b6 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 6 ), align 2 473 %b7 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 7 ), align 2 474 %b8 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8 ), align 2 475 %b9 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 9 ), align 2 476 %b10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 10), align 2 477 %b11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 11), align 2 478 %b12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 12), align 2 479 %b13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 13), align 2 480 %b14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 14), align 2 481 %b15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 15), align 2 482 %b16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2 483 %b17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 17), align 2 484 %b18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 18), align 2 485 %b19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 19), align 2 486 %b20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 20), align 2 487 %b21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 21), align 2 488 %b22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 22), align 2 489 %b23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 23), align 2 490 %b24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2 491 %b25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 25), align 2 492 %b26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 26), align 2 493 %b27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 27), align 2 494 %b28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 28), align 2 495 %b29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 29), align 2 496 %b30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 30), align 2 497 %b31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 31), align 2 498 %c0 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 0 ), align 2 499 %c1 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 1 ), align 2 500 %c2 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 2 ), align 2 501 %c3 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 3 ), align 2 502 %c4 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 4 ), align 2 503 %c5 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 5 ), align 2 504 %c6 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 6 ), align 2 505 %c7 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 7 ), align 2 506 %c8 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8 ), align 2 507 %c9 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 9 ), align 2 508 %c10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 10), align 2 509 %c11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 11), align 2 510 %c12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 12), align 2 511 %c13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 13), align 2 512 %c14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 14), align 2 513 %c15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 15), align 2 514 %c16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2 515 %c17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 17), align 2 516 %c18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 18), align 2 517 %c19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 19), align 2 518 %c20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 20), align 2 519 %c21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 21), align 2 520 %c22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 22), align 2 521 %c23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 23), align 2 522 %c24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2 523 %c25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 25), align 2 524 %c26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 26), align 2 525 %c27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 27), align 2 526 %c28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 28), align 2 527 %c29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 29), align 2 528 %c30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 30), align 2 529 %c31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 31), align 2 530 %r0 = call i16 @llvm.fshl.i16(i16 %a0 , i16 %b0 , i16 %c0 ) 531 %r1 = call i16 @llvm.fshl.i16(i16 %a1 , i16 %b1 , i16 %c1 ) 532 %r2 = call i16 @llvm.fshl.i16(i16 %a2 , i16 %b2 , i16 %c2 ) 533 %r3 = call i16 @llvm.fshl.i16(i16 %a3 , i16 %b3 , i16 %c3 ) 534 %r4 = call i16 @llvm.fshl.i16(i16 %a4 , i16 %b4 , i16 %c4 ) 535 %r5 = call i16 @llvm.fshl.i16(i16 %a5 , i16 %b5 , i16 %c5 ) 536 %r6 = call i16 @llvm.fshl.i16(i16 %a6 , i16 %b6 , i16 %c6 ) 537 %r7 = call i16 @llvm.fshl.i16(i16 %a7 , i16 %b7 , i16 %c7 ) 538 %r8 = call i16 @llvm.fshl.i16(i16 %a8 , i16 %b8 , i16 %c8 ) 539 %r9 = call i16 @llvm.fshl.i16(i16 %a9 , i16 %b9 , i16 %c9 ) 540 %r10 = call i16 @llvm.fshl.i16(i16 %a10, i16 %b10, i16 %c10) 541 %r11 = call i16 @llvm.fshl.i16(i16 %a11, i16 %b11, i16 %c11) 542 %r12 = call i16 @llvm.fshl.i16(i16 %a12, i16 %b12, i16 %c12) 543 %r13 = call i16 @llvm.fshl.i16(i16 %a13, i16 %b13, i16 %c13) 544 %r14 = call i16 @llvm.fshl.i16(i16 %a14, i16 %b14, i16 %c14) 545 %r15 = call i16 @llvm.fshl.i16(i16 %a15, i16 %b15, i16 %c15) 546 %r16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 %c16) 547 %r17 = call i16 @llvm.fshl.i16(i16 %a17, i16 %b17, i16 %c17) 548 %r18 = call i16 @llvm.fshl.i16(i16 %a18, i16 %b18, i16 %c18) 549 %r19 = call i16 @llvm.fshl.i16(i16 %a19, i16 %b19, i16 %c19) 550 %r20 = call i16 @llvm.fshl.i16(i16 %a20, i16 %b20, i16 %c20) 551 %r21 = call i16 @llvm.fshl.i16(i16 %a21, i16 %b21, i16 %c21) 552 %r22 = call i16 @llvm.fshl.i16(i16 %a22, i16 %b22, i16 %c22) 553 %r23 = call i16 @llvm.fshl.i16(i16 %a23, i16 %b23, i16 %c23) 554 %r24 = call i16 @llvm.fshl.i16(i16 %a24, i16 %b24, i16 %c24) 555 %r25 = call i16 @llvm.fshl.i16(i16 %a25, i16 %b25, i16 %c25) 556 %r26 = call i16 @llvm.fshl.i16(i16 %a26, i16 %b26, i16 %c26) 557 %r27 = call i16 @llvm.fshl.i16(i16 %a27, i16 %b27, i16 %c27) 558 %r28 = call i16 @llvm.fshl.i16(i16 %a28, i16 %b28, i16 %c28) 559 %r29 = call i16 @llvm.fshl.i16(i16 %a29, i16 %b29, i16 %c29) 560 %r30 = call i16 @llvm.fshl.i16(i16 %a30, i16 %b30, i16 %c30) 561 %r31 = call i16 @llvm.fshl.i16(i16 %a31, i16 %b31, i16 %c31) 562 store i16 %r0 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 0 ), align 2 563 store i16 %r1 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 1 ), align 2 564 store i16 %r2 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 2 ), align 2 565 store i16 %r3 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 3 ), align 2 566 store i16 %r4 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 4 ), align 2 567 store i16 %r5 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 5 ), align 2 568 store i16 %r6 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 6 ), align 2 569 store i16 %r7 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 7 ), align 2 570 store i16 %r8 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 8 ), align 2 571 store i16 %r9 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 9 ), align 2 572 store i16 %r10, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 10), align 2 573 store i16 %r11, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 11), align 2 574 store i16 %r12, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 12), align 2 575 store i16 %r13, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 13), align 2 576 store i16 %r14, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 14), align 2 577 store i16 %r15, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 15), align 2 578 store i16 %r16, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 16), align 2 579 store i16 %r17, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 17), align 2 580 store i16 %r18, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 18), align 2 581 store i16 %r19, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 19), align 2 582 store i16 %r20, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 20), align 2 583 store i16 %r21, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 21), align 2 584 store i16 %r22, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 22), align 2 585 store i16 %r23, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 23), align 2 586 store i16 %r24, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 24), align 2 587 store i16 %r25, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 25), align 2 588 store i16 %r26, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 26), align 2 589 store i16 %r27, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 27), align 2 590 store i16 %r28, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 28), align 2 591 store i16 %r29, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 29), align 2 592 store i16 %r30, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 30), align 2 593 store i16 %r31, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 31), align 2 594 ret void 595} 596 597define void @fshl_v64i8() { 598; SSE-LABEL: @fshl_v64i8( 599; SSE-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr @a8, align 1 600; SSE-NEXT: [[TMP2:%.*]] = load <16 x i8>, ptr @b8, align 1 601; SSE-NEXT: [[TMP3:%.*]] = load <16 x i8>, ptr @c8, align 1 602; SSE-NEXT: [[TMP4:%.*]] = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> [[TMP1]], <16 x i8> [[TMP2]], <16 x i8> [[TMP3]]) 603; SSE-NEXT: store <16 x i8> [[TMP4]], ptr @d8, align 1 604; SSE-NEXT: [[TMP5:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1 605; SSE-NEXT: [[TMP6:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1 606; SSE-NEXT: [[TMP7:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1 607; SSE-NEXT: [[TMP8:%.*]] = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> [[TMP5]], <16 x i8> [[TMP6]], <16 x i8> [[TMP7]]) 608; SSE-NEXT: store <16 x i8> [[TMP8]], ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 16), align 1 609; SSE-NEXT: [[TMP9:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1 610; SSE-NEXT: [[TMP10:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1 611; SSE-NEXT: [[TMP11:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1 612; SSE-NEXT: [[TMP12:%.*]] = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> [[TMP9]], <16 x i8> [[TMP10]], <16 x i8> [[TMP11]]) 613; SSE-NEXT: store <16 x i8> [[TMP12]], ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 32), align 1 614; SSE-NEXT: [[TMP13:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1 615; SSE-NEXT: [[TMP14:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1 616; SSE-NEXT: [[TMP15:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1 617; SSE-NEXT: [[TMP16:%.*]] = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> [[TMP13]], <16 x i8> [[TMP14]], <16 x i8> [[TMP15]]) 618; SSE-NEXT: store <16 x i8> [[TMP16]], ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 48), align 1 619; SSE-NEXT: ret void 620; 621; AVX-LABEL: @fshl_v64i8( 622; AVX-NEXT: [[TMP1:%.*]] = load <32 x i8>, ptr @a8, align 1 623; AVX-NEXT: [[TMP2:%.*]] = load <32 x i8>, ptr @b8, align 1 624; AVX-NEXT: [[TMP3:%.*]] = load <32 x i8>, ptr @c8, align 1 625; AVX-NEXT: [[TMP4:%.*]] = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> [[TMP1]], <32 x i8> [[TMP2]], <32 x i8> [[TMP3]]) 626; AVX-NEXT: store <32 x i8> [[TMP4]], ptr @d8, align 1 627; AVX-NEXT: [[TMP5:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1 628; AVX-NEXT: [[TMP6:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1 629; AVX-NEXT: [[TMP7:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1 630; AVX-NEXT: [[TMP8:%.*]] = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> [[TMP5]], <32 x i8> [[TMP6]], <32 x i8> [[TMP7]]) 631; AVX-NEXT: store <32 x i8> [[TMP8]], ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 32), align 1 632; AVX-NEXT: ret void 633; 634; AVX512-LABEL: @fshl_v64i8( 635; AVX512-NEXT: [[TMP1:%.*]] = load <64 x i8>, ptr @a8, align 1 636; AVX512-NEXT: [[TMP2:%.*]] = load <64 x i8>, ptr @b8, align 1 637; AVX512-NEXT: [[TMP3:%.*]] = load <64 x i8>, ptr @c8, align 1 638; AVX512-NEXT: [[TMP4:%.*]] = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> [[TMP1]], <64 x i8> [[TMP2]], <64 x i8> [[TMP3]]) 639; AVX512-NEXT: store <64 x i8> [[TMP4]], ptr @d8, align 1 640; AVX512-NEXT: ret void 641; 642; AVX512VBMI2-LABEL: @fshl_v64i8( 643; AVX512VBMI2-NEXT: [[TMP1:%.*]] = load <64 x i8>, ptr @a8, align 1 644; AVX512VBMI2-NEXT: [[TMP2:%.*]] = load <64 x i8>, ptr @b8, align 1 645; AVX512VBMI2-NEXT: [[TMP3:%.*]] = load <64 x i8>, ptr @c8, align 1 646; AVX512VBMI2-NEXT: [[TMP4:%.*]] = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> [[TMP1]], <64 x i8> [[TMP2]], <64 x i8> [[TMP3]]) 647; AVX512VBMI2-NEXT: store <64 x i8> [[TMP4]], ptr @d8, align 1 648; AVX512VBMI2-NEXT: ret void 649; 650 %a0 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 0 ), align 1 651 %a1 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 1 ), align 1 652 %a2 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 2 ), align 1 653 %a3 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 3 ), align 1 654 %a4 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 4 ), align 1 655 %a5 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 5 ), align 1 656 %a6 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 6 ), align 1 657 %a7 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 7 ), align 1 658 %a8 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 8 ), align 1 659 %a9 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 9 ), align 1 660 %a10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 10), align 1 661 %a11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 11), align 1 662 %a12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 12), align 1 663 %a13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 13), align 1 664 %a14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 14), align 1 665 %a15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 15), align 1 666 %a16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1 667 %a17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 17), align 1 668 %a18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 18), align 1 669 %a19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 19), align 1 670 %a20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 20), align 1 671 %a21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 21), align 1 672 %a22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 22), align 1 673 %a23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 23), align 1 674 %a24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 24), align 1 675 %a25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 25), align 1 676 %a26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 26), align 1 677 %a27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 27), align 1 678 %a28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 28), align 1 679 %a29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 29), align 1 680 %a30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 30), align 1 681 %a31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 31), align 1 682 %a32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1 683 %a33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 33), align 1 684 %a34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 34), align 1 685 %a35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 35), align 1 686 %a36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 36), align 1 687 %a37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 37), align 1 688 %a38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 38), align 1 689 %a39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 39), align 1 690 %a40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 40), align 1 691 %a41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 41), align 1 692 %a42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 42), align 1 693 %a43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 43), align 1 694 %a44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 44), align 1 695 %a45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 45), align 1 696 %a46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 46), align 1 697 %a47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 47), align 1 698 %a48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1 699 %a49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 49), align 1 700 %a50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 50), align 1 701 %a51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 51), align 1 702 %a52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 52), align 1 703 %a53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 53), align 1 704 %a54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 54), align 1 705 %a55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 55), align 1 706 %a56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 56), align 1 707 %a57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 57), align 1 708 %a58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 58), align 1 709 %a59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 59), align 1 710 %a60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 60), align 1 711 %a61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 61), align 1 712 %a62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 62), align 1 713 %a63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 63), align 1 714 %b0 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 0 ), align 1 715 %b1 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 1 ), align 1 716 %b2 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 2 ), align 1 717 %b3 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 3 ), align 1 718 %b4 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 4 ), align 1 719 %b5 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 5 ), align 1 720 %b6 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 6 ), align 1 721 %b7 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 7 ), align 1 722 %b8 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 8 ), align 1 723 %b9 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 9 ), align 1 724 %b10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 10), align 1 725 %b11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 11), align 1 726 %b12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 12), align 1 727 %b13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 13), align 1 728 %b14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 14), align 1 729 %b15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 15), align 1 730 %b16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1 731 %b17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 17), align 1 732 %b18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 18), align 1 733 %b19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 19), align 1 734 %b20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 20), align 1 735 %b21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 21), align 1 736 %b22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 22), align 1 737 %b23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 23), align 1 738 %b24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 24), align 1 739 %b25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 25), align 1 740 %b26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 26), align 1 741 %b27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 27), align 1 742 %b28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 28), align 1 743 %b29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 29), align 1 744 %b30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 30), align 1 745 %b31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 31), align 1 746 %b32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1 747 %b33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 33), align 1 748 %b34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 34), align 1 749 %b35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 35), align 1 750 %b36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 36), align 1 751 %b37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 37), align 1 752 %b38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 38), align 1 753 %b39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 39), align 1 754 %b40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 40), align 1 755 %b41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 41), align 1 756 %b42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 42), align 1 757 %b43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 43), align 1 758 %b44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 44), align 1 759 %b45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 45), align 1 760 %b46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 46), align 1 761 %b47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 47), align 1 762 %b48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1 763 %b49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 49), align 1 764 %b50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 50), align 1 765 %b51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 51), align 1 766 %b52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 52), align 1 767 %b53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 53), align 1 768 %b54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 54), align 1 769 %b55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 55), align 1 770 %b56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 56), align 1 771 %b57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 57), align 1 772 %b58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 58), align 1 773 %b59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 59), align 1 774 %b60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 60), align 1 775 %b61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 61), align 1 776 %b62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 62), align 1 777 %b63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 63), align 1 778 %c0 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 0 ), align 1 779 %c1 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 1 ), align 1 780 %c2 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 2 ), align 1 781 %c3 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 3 ), align 1 782 %c4 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 4 ), align 1 783 %c5 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 5 ), align 1 784 %c6 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 6 ), align 1 785 %c7 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 7 ), align 1 786 %c8 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 8 ), align 1 787 %c9 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 9 ), align 1 788 %c10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 10), align 1 789 %c11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 11), align 1 790 %c12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 12), align 1 791 %c13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 13), align 1 792 %c14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 14), align 1 793 %c15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 15), align 1 794 %c16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1 795 %c17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 17), align 1 796 %c18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 18), align 1 797 %c19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 19), align 1 798 %c20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 20), align 1 799 %c21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 21), align 1 800 %c22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 22), align 1 801 %c23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 23), align 1 802 %c24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 24), align 1 803 %c25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 25), align 1 804 %c26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 26), align 1 805 %c27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 27), align 1 806 %c28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 28), align 1 807 %c29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 29), align 1 808 %c30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 30), align 1 809 %c31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 31), align 1 810 %c32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1 811 %c33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 33), align 1 812 %c34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 34), align 1 813 %c35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 35), align 1 814 %c36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 36), align 1 815 %c37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 37), align 1 816 %c38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 38), align 1 817 %c39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 39), align 1 818 %c40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 40), align 1 819 %c41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 41), align 1 820 %c42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 42), align 1 821 %c43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 43), align 1 822 %c44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 44), align 1 823 %c45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 45), align 1 824 %c46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 46), align 1 825 %c47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 47), align 1 826 %c48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1 827 %c49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 49), align 1 828 %c50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 50), align 1 829 %c51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 51), align 1 830 %c52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 52), align 1 831 %c53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 53), align 1 832 %c54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 54), align 1 833 %c55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 55), align 1 834 %c56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 56), align 1 835 %c57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 57), align 1 836 %c58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 58), align 1 837 %c59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 59), align 1 838 %c60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 60), align 1 839 %c61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 61), align 1 840 %c62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 62), align 1 841 %c63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 63), align 1 842 %r0 = call i8 @llvm.fshl.i8(i8 %a0 , i8 %b0 , i8 %c0 ) 843 %r1 = call i8 @llvm.fshl.i8(i8 %a1 , i8 %b1 , i8 %c1 ) 844 %r2 = call i8 @llvm.fshl.i8(i8 %a2 , i8 %b2 , i8 %c2 ) 845 %r3 = call i8 @llvm.fshl.i8(i8 %a3 , i8 %b3 , i8 %c3 ) 846 %r4 = call i8 @llvm.fshl.i8(i8 %a4 , i8 %b4 , i8 %c4 ) 847 %r5 = call i8 @llvm.fshl.i8(i8 %a5 , i8 %b5 , i8 %c5 ) 848 %r6 = call i8 @llvm.fshl.i8(i8 %a6 , i8 %b6 , i8 %c6 ) 849 %r7 = call i8 @llvm.fshl.i8(i8 %a7 , i8 %b7 , i8 %c7 ) 850 %r8 = call i8 @llvm.fshl.i8(i8 %a8 , i8 %b8 , i8 %c8 ) 851 %r9 = call i8 @llvm.fshl.i8(i8 %a9 , i8 %b9 , i8 %c9 ) 852 %r10 = call i8 @llvm.fshl.i8(i8 %a10, i8 %b10, i8 %c10) 853 %r11 = call i8 @llvm.fshl.i8(i8 %a11, i8 %b11, i8 %c11) 854 %r12 = call i8 @llvm.fshl.i8(i8 %a12, i8 %b12, i8 %c12) 855 %r13 = call i8 @llvm.fshl.i8(i8 %a13, i8 %b13, i8 %c13) 856 %r14 = call i8 @llvm.fshl.i8(i8 %a14, i8 %b14, i8 %c14) 857 %r15 = call i8 @llvm.fshl.i8(i8 %a15, i8 %b15, i8 %c15) 858 %r16 = call i8 @llvm.fshl.i8(i8 %a16, i8 %b16, i8 %c16) 859 %r17 = call i8 @llvm.fshl.i8(i8 %a17, i8 %b17, i8 %c17) 860 %r18 = call i8 @llvm.fshl.i8(i8 %a18, i8 %b18, i8 %c18) 861 %r19 = call i8 @llvm.fshl.i8(i8 %a19, i8 %b19, i8 %c19) 862 %r20 = call i8 @llvm.fshl.i8(i8 %a20, i8 %b20, i8 %c20) 863 %r21 = call i8 @llvm.fshl.i8(i8 %a21, i8 %b21, i8 %c21) 864 %r22 = call i8 @llvm.fshl.i8(i8 %a22, i8 %b22, i8 %c22) 865 %r23 = call i8 @llvm.fshl.i8(i8 %a23, i8 %b23, i8 %c23) 866 %r24 = call i8 @llvm.fshl.i8(i8 %a24, i8 %b24, i8 %c24) 867 %r25 = call i8 @llvm.fshl.i8(i8 %a25, i8 %b25, i8 %c25) 868 %r26 = call i8 @llvm.fshl.i8(i8 %a26, i8 %b26, i8 %c26) 869 %r27 = call i8 @llvm.fshl.i8(i8 %a27, i8 %b27, i8 %c27) 870 %r28 = call i8 @llvm.fshl.i8(i8 %a28, i8 %b28, i8 %c28) 871 %r29 = call i8 @llvm.fshl.i8(i8 %a29, i8 %b29, i8 %c29) 872 %r30 = call i8 @llvm.fshl.i8(i8 %a30, i8 %b30, i8 %c30) 873 %r31 = call i8 @llvm.fshl.i8(i8 %a31, i8 %b31, i8 %c31) 874 %r32 = call i8 @llvm.fshl.i8(i8 %a32, i8 %b32, i8 %c32) 875 %r33 = call i8 @llvm.fshl.i8(i8 %a33, i8 %b33, i8 %c33) 876 %r34 = call i8 @llvm.fshl.i8(i8 %a34, i8 %b34, i8 %c34) 877 %r35 = call i8 @llvm.fshl.i8(i8 %a35, i8 %b35, i8 %c35) 878 %r36 = call i8 @llvm.fshl.i8(i8 %a36, i8 %b36, i8 %c36) 879 %r37 = call i8 @llvm.fshl.i8(i8 %a37, i8 %b37, i8 %c37) 880 %r38 = call i8 @llvm.fshl.i8(i8 %a38, i8 %b38, i8 %c38) 881 %r39 = call i8 @llvm.fshl.i8(i8 %a39, i8 %b39, i8 %c39) 882 %r40 = call i8 @llvm.fshl.i8(i8 %a40, i8 %b40, i8 %c40) 883 %r41 = call i8 @llvm.fshl.i8(i8 %a41, i8 %b41, i8 %c41) 884 %r42 = call i8 @llvm.fshl.i8(i8 %a42, i8 %b42, i8 %c42) 885 %r43 = call i8 @llvm.fshl.i8(i8 %a43, i8 %b43, i8 %c43) 886 %r44 = call i8 @llvm.fshl.i8(i8 %a44, i8 %b44, i8 %c44) 887 %r45 = call i8 @llvm.fshl.i8(i8 %a45, i8 %b45, i8 %c45) 888 %r46 = call i8 @llvm.fshl.i8(i8 %a46, i8 %b46, i8 %c46) 889 %r47 = call i8 @llvm.fshl.i8(i8 %a47, i8 %b47, i8 %c47) 890 %r48 = call i8 @llvm.fshl.i8(i8 %a48, i8 %b48, i8 %c48) 891 %r49 = call i8 @llvm.fshl.i8(i8 %a49, i8 %b49, i8 %c49) 892 %r50 = call i8 @llvm.fshl.i8(i8 %a50, i8 %b50, i8 %c50) 893 %r51 = call i8 @llvm.fshl.i8(i8 %a51, i8 %b51, i8 %c51) 894 %r52 = call i8 @llvm.fshl.i8(i8 %a52, i8 %b52, i8 %c52) 895 %r53 = call i8 @llvm.fshl.i8(i8 %a53, i8 %b53, i8 %c53) 896 %r54 = call i8 @llvm.fshl.i8(i8 %a54, i8 %b54, i8 %c54) 897 %r55 = call i8 @llvm.fshl.i8(i8 %a55, i8 %b55, i8 %c55) 898 %r56 = call i8 @llvm.fshl.i8(i8 %a56, i8 %b56, i8 %c56) 899 %r57 = call i8 @llvm.fshl.i8(i8 %a57, i8 %b57, i8 %c57) 900 %r58 = call i8 @llvm.fshl.i8(i8 %a58, i8 %b58, i8 %c58) 901 %r59 = call i8 @llvm.fshl.i8(i8 %a59, i8 %b59, i8 %c59) 902 %r60 = call i8 @llvm.fshl.i8(i8 %a60, i8 %b60, i8 %c60) 903 %r61 = call i8 @llvm.fshl.i8(i8 %a61, i8 %b61, i8 %c61) 904 %r62 = call i8 @llvm.fshl.i8(i8 %a62, i8 %b62, i8 %c62) 905 %r63 = call i8 @llvm.fshl.i8(i8 %a63, i8 %b63, i8 %c63) 906 store i8 %r0 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 0 ), align 1 907 store i8 %r1 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 1 ), align 1 908 store i8 %r2 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 2 ), align 1 909 store i8 %r3 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 3 ), align 1 910 store i8 %r4 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 4 ), align 1 911 store i8 %r5 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 5 ), align 1 912 store i8 %r6 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 6 ), align 1 913 store i8 %r7 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 7 ), align 1 914 store i8 %r8 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 8 ), align 1 915 store i8 %r9 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 9 ), align 1 916 store i8 %r10, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 10), align 1 917 store i8 %r11, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 11), align 1 918 store i8 %r12, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 12), align 1 919 store i8 %r13, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 13), align 1 920 store i8 %r14, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 14), align 1 921 store i8 %r15, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 15), align 1 922 store i8 %r16, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 16), align 1 923 store i8 %r17, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 17), align 1 924 store i8 %r18, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 18), align 1 925 store i8 %r19, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 19), align 1 926 store i8 %r20, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 20), align 1 927 store i8 %r21, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 21), align 1 928 store i8 %r22, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 22), align 1 929 store i8 %r23, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 23), align 1 930 store i8 %r24, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 24), align 1 931 store i8 %r25, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 25), align 1 932 store i8 %r26, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 26), align 1 933 store i8 %r27, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 27), align 1 934 store i8 %r28, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 28), align 1 935 store i8 %r29, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 29), align 1 936 store i8 %r30, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 30), align 1 937 store i8 %r31, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 31), align 1 938 store i8 %r32, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 32), align 1 939 store i8 %r33, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 33), align 1 940 store i8 %r34, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 34), align 1 941 store i8 %r35, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 35), align 1 942 store i8 %r36, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 36), align 1 943 store i8 %r37, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 37), align 1 944 store i8 %r38, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 38), align 1 945 store i8 %r39, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 39), align 1 946 store i8 %r40, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 40), align 1 947 store i8 %r41, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 41), align 1 948 store i8 %r42, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 42), align 1 949 store i8 %r43, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 43), align 1 950 store i8 %r44, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 44), align 1 951 store i8 %r45, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 45), align 1 952 store i8 %r46, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 46), align 1 953 store i8 %r47, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 47), align 1 954 store i8 %r48, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 48), align 1 955 store i8 %r49, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 49), align 1 956 store i8 %r50, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 50), align 1 957 store i8 %r51, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 51), align 1 958 store i8 %r52, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 52), align 1 959 store i8 %r53, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 53), align 1 960 store i8 %r54, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 54), align 1 961 store i8 %r55, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 55), align 1 962 store i8 %r56, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 56), align 1 963 store i8 %r57, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 57), align 1 964 store i8 %r58, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 58), align 1 965 store i8 %r59, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 59), align 1 966 store i8 %r60, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 60), align 1 967 store i8 %r61, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 61), align 1 968 store i8 %r62, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 62), align 1 969 store i8 %r63, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 63), align 1 970 ret void 971} 972