brintos

brintos / llvm-project-archived public Read only

0
0
Text · 79.0 KiB · 217f5e0 Raw
972 lines · plain
1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py2; RUN: opt < %s -mtriple=x86_64-unknown -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=SSE3; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=slm -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=SSE4; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7-avx -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX,AVX15; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=core-avx2 -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX,AVX26; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=skx -mattr=+prefer-256-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX,AVX2567; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=skx -mattr=-prefer-256-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX5128; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=znver4 -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX512VBMI29; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=knl -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX51210 11@a64 = common global [8 x i64] zeroinitializer, align 6412@b64 = common global [8 x i64] zeroinitializer, align 6413@c64 = common global [8 x i64] zeroinitializer, align 6414@d64 = common global [8 x i64] zeroinitializer, align 6415@a32 = common global [16 x i32] zeroinitializer, align 6416@b32 = common global [16 x i32] zeroinitializer, align 6417@c32 = common global [16 x i32] zeroinitializer, align 6418@d32 = common global [16 x i32] zeroinitializer, align 6419@a16 = common global [32 x i16] zeroinitializer, align 6420@b16 = common global [32 x i16] zeroinitializer, align 6421@c16 = common global [32 x i16] zeroinitializer, align 6422@d16 = common global [32 x i16] zeroinitializer, align 6423@a8  = common global [64 x i8] zeroinitializer, align 6424@b8  = common global [64 x i8] zeroinitializer, align 6425@c8  = common global [64 x i8] zeroinitializer, align 6426@d8  = common global [64 x i8] zeroinitializer, align 6427 28declare i64 @llvm.fshr.i64(i64, i64, i64)29declare i32 @llvm.fshr.i32(i32, i32, i32)30declare i16 @llvm.fshr.i16(i16, i16, i16)31declare i8  @llvm.fshr.i8 (i8 , i8 , i8 )32 33define void @fshr_v8i64() {34; SSE-LABEL: @fshr_v8i64(35; SSE-NEXT:    [[A0:%.*]] = load i64, ptr @a64, align 836; SSE-NEXT:    [[A1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 837; SSE-NEXT:    [[A2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 838; SSE-NEXT:    [[A3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 839; SSE-NEXT:    [[A4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 840; SSE-NEXT:    [[A5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 841; SSE-NEXT:    [[A6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 842; SSE-NEXT:    [[A7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 843; SSE-NEXT:    [[B0:%.*]] = load i64, ptr @b64, align 844; SSE-NEXT:    [[B1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 845; SSE-NEXT:    [[B2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 846; SSE-NEXT:    [[B3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 847; SSE-NEXT:    [[B4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 848; SSE-NEXT:    [[B5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 849; SSE-NEXT:    [[B6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 850; SSE-NEXT:    [[B7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 851; SSE-NEXT:    [[C0:%.*]] = load i64, ptr @c64, align 852; SSE-NEXT:    [[C1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 1), align 853; SSE-NEXT:    [[C2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 854; SSE-NEXT:    [[C3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 3), align 855; SSE-NEXT:    [[C4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 856; SSE-NEXT:    [[C5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 5), align 857; SSE-NEXT:    [[C6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 858; SSE-NEXT:    [[C7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 7), align 859; SSE-NEXT:    [[R0:%.*]] = call i64 @llvm.fshr.i64(i64 [[A0]], i64 [[B0]], i64 [[C0]])60; SSE-NEXT:    [[R1:%.*]] = call i64 @llvm.fshr.i64(i64 [[A1]], i64 [[B1]], i64 [[C1]])61; SSE-NEXT:    [[R2:%.*]] = call i64 @llvm.fshr.i64(i64 [[A2]], i64 [[B2]], i64 [[C2]])62; SSE-NEXT:    [[R3:%.*]] = call i64 @llvm.fshr.i64(i64 [[A3]], i64 [[B3]], i64 [[C3]])63; SSE-NEXT:    [[R4:%.*]] = call i64 @llvm.fshr.i64(i64 [[A4]], i64 [[B4]], i64 [[C4]])64; SSE-NEXT:    [[R5:%.*]] = call i64 @llvm.fshr.i64(i64 [[A5]], i64 [[B5]], i64 [[C5]])65; SSE-NEXT:    [[R6:%.*]] = call i64 @llvm.fshr.i64(i64 [[A6]], i64 [[B6]], i64 [[C6]])66; SSE-NEXT:    [[R7:%.*]] = call i64 @llvm.fshr.i64(i64 [[A7]], i64 [[B7]], i64 [[C7]])67; SSE-NEXT:    store i64 [[R0]], ptr @d64, align 868; SSE-NEXT:    store i64 [[R1]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 1), align 869; SSE-NEXT:    store i64 [[R2]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 2), align 870; SSE-NEXT:    store i64 [[R3]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 3), align 871; SSE-NEXT:    store i64 [[R4]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 872; SSE-NEXT:    store i64 [[R5]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 5), align 873; SSE-NEXT:    store i64 [[R6]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 6), align 874; SSE-NEXT:    store i64 [[R7]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 7), align 875; SSE-NEXT:    ret void76;77; AVX1-LABEL: @fshr_v8i64(78; AVX1-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @a64, align 879; AVX1-NEXT:    [[TMP2:%.*]] = load <2 x i64>, ptr @b64, align 880; AVX1-NEXT:    [[TMP3:%.*]] = load <2 x i64>, ptr @c64, align 881; AVX1-NEXT:    [[TMP4:%.*]] = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> [[TMP1]], <2 x i64> [[TMP2]], <2 x i64> [[TMP3]])82; AVX1-NEXT:    store <2 x i64> [[TMP4]], ptr @d64, align 883; AVX1-NEXT:    [[TMP5:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 884; AVX1-NEXT:    [[TMP6:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 885; AVX1-NEXT:    [[TMP7:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 886; AVX1-NEXT:    [[TMP8:%.*]] = call <2 x i64> @llvm.fshr.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 888; AVX1-NEXT:    [[TMP9:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 889; AVX1-NEXT:    [[TMP10:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 890; AVX1-NEXT:    [[TMP11:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 891; AVX1-NEXT:    [[TMP12:%.*]] = call <2 x i64> @llvm.fshr.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 893; AVX1-NEXT:    [[TMP13:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 894; AVX1-NEXT:    [[TMP14:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 895; AVX1-NEXT:    [[TMP15:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 896; AVX1-NEXT:    [[TMP16:%.*]] = call <2 x i64> @llvm.fshr.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 898; AVX1-NEXT:    ret void99;100; AVX2-LABEL: @fshr_v8i64(101; AVX2-NEXT:    [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 8102; AVX2-NEXT:    [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 8103; AVX2-NEXT:    [[TMP3:%.*]] = load <4 x i64>, ptr @c64, align 8104; AVX2-NEXT:    [[TMP4:%.*]] = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> [[TMP1]], <4 x i64> [[TMP2]], <4 x i64> [[TMP3]])105; AVX2-NEXT:    store <4 x i64> [[TMP4]], ptr @d64, align 8106; AVX2-NEXT:    [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8107; AVX2-NEXT:    [[TMP6:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8108; AVX2-NEXT:    [[TMP7:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8109; AVX2-NEXT:    [[TMP8:%.*]] = call <4 x i64> @llvm.fshr.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 8111; AVX2-NEXT:    ret void112;113; AVX256-LABEL: @fshr_v8i64(114; AVX256-NEXT:    [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 8115; AVX256-NEXT:    [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 8116; AVX256-NEXT:    [[TMP3:%.*]] = load <4 x i64>, ptr @c64, align 8117; AVX256-NEXT:    [[TMP4:%.*]] = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> [[TMP1]], <4 x i64> [[TMP2]], <4 x i64> [[TMP3]])118; AVX256-NEXT:    store <4 x i64> [[TMP4]], ptr @d64, align 8119; AVX256-NEXT:    [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8120; AVX256-NEXT:    [[TMP6:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8121; AVX256-NEXT:    [[TMP7:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8122; AVX256-NEXT:    [[TMP8:%.*]] = call <4 x i64> @llvm.fshr.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 8124; AVX256-NEXT:    ret void125;126; AVX512-LABEL: @fshr_v8i64(127; AVX512-NEXT:    [[TMP1:%.*]] = load <8 x i64>, ptr @a64, align 8128; AVX512-NEXT:    [[TMP2:%.*]] = load <8 x i64>, ptr @b64, align 8129; AVX512-NEXT:    [[TMP3:%.*]] = load <8 x i64>, ptr @c64, align 8130; AVX512-NEXT:    [[TMP4:%.*]] = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> [[TMP1]], <8 x i64> [[TMP2]], <8 x i64> [[TMP3]])131; AVX512-NEXT:    store <8 x i64> [[TMP4]], ptr @d64, align 8132; AVX512-NEXT:    ret void133;134; AVX512VBMI2-LABEL: @fshr_v8i64(135; AVX512VBMI2-NEXT:    [[TMP1:%.*]] = load <8 x i64>, ptr @a64, align 8136; AVX512VBMI2-NEXT:    [[TMP2:%.*]] = load <8 x i64>, ptr @b64, align 8137; AVX512VBMI2-NEXT:    [[TMP3:%.*]] = load <8 x i64>, ptr @c64, align 8138; AVX512VBMI2-NEXT:    [[TMP4:%.*]] = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> [[TMP1]], <8 x i64> [[TMP2]], <8 x i64> [[TMP3]])139; AVX512VBMI2-NEXT:    store <8 x i64> [[TMP4]], ptr @d64, align 8140; AVX512VBMI2-NEXT:    ret void141;142  %a0 = load i64, ptr @a64, align 8143  %a1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 8144  %a2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8145  %a3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 8146  %a4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8147  %a5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 8148  %a6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8149  %a7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 8150  %b0 = load i64, ptr @b64, align 8151  %b1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 8152  %b2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8153  %b3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 8154  %b4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8155  %b5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 8156  %b6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8157  %b7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 8158  %c0 = load i64, ptr @c64, align 8159  %c1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 1), align 8160  %c2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8161  %c3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 3), align 8162  %c4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8163  %c5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 5), align 8164  %c6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8165  %c7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 7), align 8166  %r0 = call i64 @llvm.fshr.i64(i64 %a0, i64 %b0, i64 %c0)167  %r1 = call i64 @llvm.fshr.i64(i64 %a1, i64 %b1, i64 %c1)168  %r2 = call i64 @llvm.fshr.i64(i64 %a2, i64 %b2, i64 %c2)169  %r3 = call i64 @llvm.fshr.i64(i64 %a3, i64 %b3, i64 %c3)170  %r4 = call i64 @llvm.fshr.i64(i64 %a4, i64 %b4, i64 %c4)171  %r5 = call i64 @llvm.fshr.i64(i64 %a5, i64 %b5, i64 %c5)172  %r6 = call i64 @llvm.fshr.i64(i64 %a6, i64 %b6, i64 %c6)173  %r7 = call i64 @llvm.fshr.i64(i64 %a7, i64 %b7, i64 %c7)174  store i64 %r0, ptr @d64, align 8175  store i64 %r1, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 1), align 8176  store i64 %r2, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 2), align 8177  store i64 %r3, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 3), align 8178  store i64 %r4, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 8179  store i64 %r5, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 5), align 8180  store i64 %r6, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 6), align 8181  store i64 %r7, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 7), align 8182  ret void183}184 185define void @fshr_v16i32() {186; SSE-LABEL: @fshr_v16i32(187; SSE-NEXT:    [[A0:%.*]] = load i32, ptr @a32, align 4188; SSE-NEXT:    [[A1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1), align 4189; SSE-NEXT:    [[A2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2), align 4190; SSE-NEXT:    [[A3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3), align 4191; SSE-NEXT:    [[A4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4192; SSE-NEXT:    [[A5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5), align 4193; SSE-NEXT:    [[A6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6), align 4194; SSE-NEXT:    [[A7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7), align 4195; SSE-NEXT:    [[A8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4196; SSE-NEXT:    [[A9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9), align 4197; SSE-NEXT:    [[A10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 4198; SSE-NEXT:    [[A11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 4199; SSE-NEXT:    [[A12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4200; SSE-NEXT:    [[A13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 4201; SSE-NEXT:    [[A14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 4202; SSE-NEXT:    [[A15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 4203; SSE-NEXT:    [[B0:%.*]] = load i32, ptr @b32, align 4204; SSE-NEXT:    [[B1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1), align 4205; SSE-NEXT:    [[B2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2), align 4206; SSE-NEXT:    [[B3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3), align 4207; SSE-NEXT:    [[B4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4208; SSE-NEXT:    [[B5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5), align 4209; SSE-NEXT:    [[B6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6), align 4210; SSE-NEXT:    [[B7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7), align 4211; SSE-NEXT:    [[B8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4212; SSE-NEXT:    [[B9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9), align 4213; SSE-NEXT:    [[B10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 4214; SSE-NEXT:    [[B11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 4215; SSE-NEXT:    [[B12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4216; SSE-NEXT:    [[B13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 4217; SSE-NEXT:    [[B14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 4218; SSE-NEXT:    [[B15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 4219; SSE-NEXT:    [[C0:%.*]] = load i32, ptr @c32, align 4220; SSE-NEXT:    [[C1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 1), align 4221; SSE-NEXT:    [[C2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 2), align 4222; SSE-NEXT:    [[C3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 3), align 4223; SSE-NEXT:    [[C4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4), align 4224; SSE-NEXT:    [[C5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 5), align 4225; SSE-NEXT:    [[C6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 6), align 4226; SSE-NEXT:    [[C7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 7), align 4227; SSE-NEXT:    [[C8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4228; SSE-NEXT:    [[C9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 9), align 4229; SSE-NEXT:    [[C10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 10), align 4230; SSE-NEXT:    [[C11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 11), align 4231; SSE-NEXT:    [[C12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4232; SSE-NEXT:    [[C13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 13), align 4233; SSE-NEXT:    [[C14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 14), align 4234; SSE-NEXT:    [[C15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 15), align 4235; SSE-NEXT:    [[R0:%.*]] = call i32 @llvm.fshr.i32(i32 [[A0]], i32 [[B0]], i32 [[C0]])236; SSE-NEXT:    [[R1:%.*]] = call i32 @llvm.fshr.i32(i32 [[A1]], i32 [[B1]], i32 [[C1]])237; SSE-NEXT:    [[R2:%.*]] = call i32 @llvm.fshr.i32(i32 [[A2]], i32 [[B2]], i32 [[C2]])238; SSE-NEXT:    [[R3:%.*]] = call i32 @llvm.fshr.i32(i32 [[A3]], i32 [[B3]], i32 [[C3]])239; SSE-NEXT:    [[R4:%.*]] = call i32 @llvm.fshr.i32(i32 [[A4]], i32 [[B4]], i32 [[C4]])240; SSE-NEXT:    [[R5:%.*]] = call i32 @llvm.fshr.i32(i32 [[A5]], i32 [[B5]], i32 [[C5]])241; SSE-NEXT:    [[R6:%.*]] = call i32 @llvm.fshr.i32(i32 [[A6]], i32 [[B6]], i32 [[C6]])242; SSE-NEXT:    [[R7:%.*]] = call i32 @llvm.fshr.i32(i32 [[A7]], i32 [[B7]], i32 [[C7]])243; SSE-NEXT:    [[R8:%.*]] = call i32 @llvm.fshr.i32(i32 [[A8]], i32 [[B8]], i32 [[C8]])244; SSE-NEXT:    [[R9:%.*]] = call i32 @llvm.fshr.i32(i32 [[A9]], i32 [[B9]], i32 [[C9]])245; SSE-NEXT:    [[R10:%.*]] = call i32 @llvm.fshr.i32(i32 [[A10]], i32 [[B10]], i32 [[C10]])246; SSE-NEXT:    [[R11:%.*]] = call i32 @llvm.fshr.i32(i32 [[A11]], i32 [[B11]], i32 [[C11]])247; SSE-NEXT:    [[R12:%.*]] = call i32 @llvm.fshr.i32(i32 [[A12]], i32 [[B12]], i32 [[C12]])248; SSE-NEXT:    [[R13:%.*]] = call i32 @llvm.fshr.i32(i32 [[A13]], i32 [[B13]], i32 [[C13]])249; SSE-NEXT:    [[R14:%.*]] = call i32 @llvm.fshr.i32(i32 [[A14]], i32 [[B14]], i32 [[C14]])250; SSE-NEXT:    [[R15:%.*]] = call i32 @llvm.fshr.i32(i32 [[A15]], i32 [[B15]], i32 [[C15]])251; SSE-NEXT:    store i32 [[R0]], ptr @d32, align 4252; SSE-NEXT:    store i32 [[R1]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 1), align 4253; SSE-NEXT:    store i32 [[R2]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 2), align 4254; SSE-NEXT:    store i32 [[R3]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 3), align 4255; SSE-NEXT:    store i32 [[R4]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 4), align 4256; SSE-NEXT:    store i32 [[R5]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 5), align 4257; SSE-NEXT:    store i32 [[R6]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 6), align 4258; SSE-NEXT:    store i32 [[R7]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 7), align 4259; SSE-NEXT:    store i32 [[R8]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 8), align 4260; SSE-NEXT:    store i32 [[R9]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 9), align 4261; SSE-NEXT:    store i32 [[R10]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 10), align 4262; SSE-NEXT:    store i32 [[R11]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 11), align 4263; SSE-NEXT:    store i32 [[R12]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 12), align 4264; SSE-NEXT:    store i32 [[R13]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 13), align 4265; SSE-NEXT:    store i32 [[R14]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 14), align 4266; SSE-NEXT:    store i32 [[R15]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 15), align 4267; SSE-NEXT:    ret void268;269; AVX-LABEL: @fshr_v16i32(270; AVX-NEXT:    [[TMP1:%.*]] = load <8 x i32>, ptr @a32, align 4271; AVX-NEXT:    [[TMP2:%.*]] = load <8 x i32>, ptr @b32, align 4272; AVX-NEXT:    [[TMP3:%.*]] = load <8 x i32>, ptr @c32, align 4273; AVX-NEXT:    [[TMP4:%.*]] = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> [[TMP1]], <8 x i32> [[TMP2]], <8 x i32> [[TMP3]])274; AVX-NEXT:    store <8 x i32> [[TMP4]], ptr @d32, align 4275; AVX-NEXT:    [[TMP5:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4276; AVX-NEXT:    [[TMP6:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4277; AVX-NEXT:    [[TMP7:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4278; AVX-NEXT:    [[TMP8:%.*]] = call <8 x i32> @llvm.fshr.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 4280; AVX-NEXT:    ret void281;282; AVX512-LABEL: @fshr_v16i32(283; AVX512-NEXT:    [[TMP1:%.*]] = load <16 x i32>, ptr @a32, align 4284; AVX512-NEXT:    [[TMP2:%.*]] = load <16 x i32>, ptr @b32, align 4285; AVX512-NEXT:    [[TMP3:%.*]] = load <16 x i32>, ptr @c32, align 4286; AVX512-NEXT:    [[TMP4:%.*]] = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> [[TMP1]], <16 x i32> [[TMP2]], <16 x i32> [[TMP3]])287; AVX512-NEXT:    store <16 x i32> [[TMP4]], ptr @d32, align 4288; AVX512-NEXT:    ret void289;290; AVX512VBMI2-LABEL: @fshr_v16i32(291; AVX512VBMI2-NEXT:    [[TMP1:%.*]] = load <16 x i32>, ptr @a32, align 4292; AVX512VBMI2-NEXT:    [[TMP2:%.*]] = load <16 x i32>, ptr @b32, align 4293; AVX512VBMI2-NEXT:    [[TMP3:%.*]] = load <16 x i32>, ptr @c32, align 4294; AVX512VBMI2-NEXT:    [[TMP4:%.*]] = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> [[TMP1]], <16 x i32> [[TMP2]], <16 x i32> [[TMP3]])295; AVX512VBMI2-NEXT:    store <16 x i32> [[TMP4]], ptr @d32, align 4296; AVX512VBMI2-NEXT:    ret void297;298  %a0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 0 ), align 4299  %a1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1 ), align 4300  %a2  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2 ), align 4301  %a3  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3 ), align 4302  %a4  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4 ), align 4303  %a5  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5 ), align 4304  %a6  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6 ), align 4305  %a7  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7 ), align 4306  %a8  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8 ), align 4307  %a9  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9 ), align 4308  %a10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 4309  %a11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 4310  %a12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4311  %a13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 4312  %a14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 4313  %a15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 4314  %b0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 0 ), align 4315  %b1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1 ), align 4316  %b2  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2 ), align 4317  %b3  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3 ), align 4318  %b4  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4 ), align 4319  %b5  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5 ), align 4320  %b6  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6 ), align 4321  %b7  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7 ), align 4322  %b8  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8 ), align 4323  %b9  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9 ), align 4324  %b10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 4325  %b11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 4326  %b12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4327  %b13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 4328  %b14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 4329  %b15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 4330  %c0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 0 ), align 4331  %c1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 1 ), align 4332  %c2  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 2 ), align 4333  %c3  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 3 ), align 4334  %c4  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4 ), align 4335  %c5  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 5 ), align 4336  %c6  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 6 ), align 4337  %c7  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 7 ), align 4338  %c8  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8 ), align 4339  %c9  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 9 ), align 4340  %c10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 10), align 4341  %c11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 11), align 4342  %c12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4343  %c13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 13), align 4344  %c14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 14), align 4345  %c15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 15), align 4346  %r0  = call i32 @llvm.fshr.i32(i32 %a0 , i32 %b0 , i32 %c0 )347  %r1  = call i32 @llvm.fshr.i32(i32 %a1 , i32 %b1 , i32 %c1 )348  %r2  = call i32 @llvm.fshr.i32(i32 %a2 , i32 %b2 , i32 %c2 )349  %r3  = call i32 @llvm.fshr.i32(i32 %a3 , i32 %b3 , i32 %c3 )350  %r4  = call i32 @llvm.fshr.i32(i32 %a4 , i32 %b4 , i32 %c4 )351  %r5  = call i32 @llvm.fshr.i32(i32 %a5 , i32 %b5 , i32 %c5 )352  %r6  = call i32 @llvm.fshr.i32(i32 %a6 , i32 %b6 , i32 %c6 )353  %r7  = call i32 @llvm.fshr.i32(i32 %a7 , i32 %b7 , i32 %c7 )354  %r8  = call i32 @llvm.fshr.i32(i32 %a8 , i32 %b8 , i32 %c8 )355  %r9  = call i32 @llvm.fshr.i32(i32 %a9 , i32 %b9 , i32 %c9 )356  %r10 = call i32 @llvm.fshr.i32(i32 %a10, i32 %b10, i32 %c10)357  %r11 = call i32 @llvm.fshr.i32(i32 %a11, i32 %b11, i32 %c11)358  %r12 = call i32 @llvm.fshr.i32(i32 %a12, i32 %b12, i32 %c12)359  %r13 = call i32 @llvm.fshr.i32(i32 %a13, i32 %b13, i32 %c13)360  %r14 = call i32 @llvm.fshr.i32(i32 %a14, i32 %b14, i32 %c14)361  %r15 = call i32 @llvm.fshr.i32(i32 %a15, i32 %b15, i32 %c15)362  store i32 %r0 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 0 ), align 4363  store i32 %r1 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 1 ), align 4364  store i32 %r2 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 2 ), align 4365  store i32 %r3 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 3 ), align 4366  store i32 %r4 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 4 ), align 4367  store i32 %r5 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 5 ), align 4368  store i32 %r6 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 6 ), align 4369  store i32 %r7 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 7 ), align 4370  store i32 %r8 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 8 ), align 4371  store i32 %r9 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 9 ), align 4372  store i32 %r10, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 10), align 4373  store i32 %r11, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 11), align 4374  store i32 %r12, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 12), align 4375  store i32 %r13, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 13), align 4376  store i32 %r14, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 14), align 4377  store i32 %r15, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 15), align 4378  ret void379}380 381define void @fshr_v32i16() {382; SSE-LABEL: @fshr_v32i16(383; SSE-NEXT:    [[TMP1:%.*]] = load <8 x i16>, ptr @a16, align 2384; SSE-NEXT:    [[TMP2:%.*]] = load <8 x i16>, ptr @b16, align 2385; SSE-NEXT:    [[TMP3:%.*]] = load <8 x i16>, ptr @c16, align 2386; SSE-NEXT:    [[TMP4:%.*]] = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> [[TMP1]], <8 x i16> [[TMP2]], <8 x i16> [[TMP3]])387; SSE-NEXT:    store <8 x i16> [[TMP4]], ptr @d16, align 2388; SSE-NEXT:    [[TMP5:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8), align 2389; SSE-NEXT:    [[TMP6:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8), align 2390; SSE-NEXT:    [[TMP7:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8), align 2391; SSE-NEXT:    [[TMP8:%.*]] = call <8 x i16> @llvm.fshr.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 2393; SSE-NEXT:    [[TMP9:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2394; SSE-NEXT:    [[TMP10:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2395; SSE-NEXT:    [[TMP11:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2396; SSE-NEXT:    [[TMP12:%.*]] = call <8 x i16> @llvm.fshr.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 2398; SSE-NEXT:    [[TMP13:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2399; SSE-NEXT:    [[TMP14:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2400; SSE-NEXT:    [[TMP15:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2401; SSE-NEXT:    [[TMP16:%.*]] = call <8 x i16> @llvm.fshr.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 2403; SSE-NEXT:    ret void404;405; AVX-LABEL: @fshr_v32i16(406; AVX-NEXT:    [[TMP1:%.*]] = load <16 x i16>, ptr @a16, align 2407; AVX-NEXT:    [[TMP2:%.*]] = load <16 x i16>, ptr @b16, align 2408; AVX-NEXT:    [[TMP3:%.*]] = load <16 x i16>, ptr @c16, align 2409; AVX-NEXT:    [[TMP4:%.*]] = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> [[TMP1]], <16 x i16> [[TMP2]], <16 x i16> [[TMP3]])410; AVX-NEXT:    store <16 x i16> [[TMP4]], ptr @d16, align 2411; AVX-NEXT:    [[TMP5:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2412; AVX-NEXT:    [[TMP6:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2413; AVX-NEXT:    [[TMP7:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2414; AVX-NEXT:    [[TMP8:%.*]] = call <16 x i16> @llvm.fshr.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 2416; AVX-NEXT:    ret void417;418; AVX512-LABEL: @fshr_v32i16(419; AVX512-NEXT:    [[TMP1:%.*]] = load <32 x i16>, ptr @a16, align 2420; AVX512-NEXT:    [[TMP2:%.*]] = load <32 x i16>, ptr @b16, align 2421; AVX512-NEXT:    [[TMP3:%.*]] = load <32 x i16>, ptr @c16, align 2422; AVX512-NEXT:    [[TMP4:%.*]] = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> [[TMP1]], <32 x i16> [[TMP2]], <32 x i16> [[TMP3]])423; AVX512-NEXT:    store <32 x i16> [[TMP4]], ptr @d16, align 2424; AVX512-NEXT:    ret void425;426; AVX512VBMI2-LABEL: @fshr_v32i16(427; AVX512VBMI2-NEXT:    [[TMP1:%.*]] = load <32 x i16>, ptr @a16, align 2428; AVX512VBMI2-NEXT:    [[TMP2:%.*]] = load <32 x i16>, ptr @b16, align 2429; AVX512VBMI2-NEXT:    [[TMP3:%.*]] = load <32 x i16>, ptr @c16, align 2430; AVX512VBMI2-NEXT:    [[TMP4:%.*]] = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> [[TMP1]], <32 x i16> [[TMP2]], <32 x i16> [[TMP3]])431; AVX512VBMI2-NEXT:    store <32 x i16> [[TMP4]], ptr @d16, align 2432; AVX512VBMI2-NEXT:    ret void433;434  %a0  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 0 ), align 2435  %a1  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 1 ), align 2436  %a2  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 2 ), align 2437  %a3  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 3 ), align 2438  %a4  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 4 ), align 2439  %a5  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 5 ), align 2440  %a6  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 6 ), align 2441  %a7  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 7 ), align 2442  %a8  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8 ), align 2443  %a9  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 9 ), align 2444  %a10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 10), align 2445  %a11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 11), align 2446  %a12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 12), align 2447  %a13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 13), align 2448  %a14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 14), align 2449  %a15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 15), align 2450  %a16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2451  %a17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 17), align 2452  %a18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 18), align 2453  %a19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 19), align 2454  %a20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 20), align 2455  %a21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 21), align 2456  %a22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 22), align 2457  %a23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 23), align 2458  %a24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2459  %a25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 25), align 2460  %a26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 26), align 2461  %a27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 27), align 2462  %a28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 28), align 2463  %a29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 29), align 2464  %a30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 30), align 2465  %a31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 31), align 2466  %b0  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 0 ), align 2467  %b1  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 1 ), align 2468  %b2  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 2 ), align 2469  %b3  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 3 ), align 2470  %b4  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 4 ), align 2471  %b5  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 5 ), align 2472  %b6  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 6 ), align 2473  %b7  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 7 ), align 2474  %b8  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8 ), align 2475  %b9  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 9 ), align 2476  %b10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 10), align 2477  %b11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 11), align 2478  %b12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 12), align 2479  %b13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 13), align 2480  %b14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 14), align 2481  %b15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 15), align 2482  %b16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2483  %b17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 17), align 2484  %b18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 18), align 2485  %b19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 19), align 2486  %b20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 20), align 2487  %b21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 21), align 2488  %b22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 22), align 2489  %b23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 23), align 2490  %b24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2491  %b25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 25), align 2492  %b26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 26), align 2493  %b27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 27), align 2494  %b28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 28), align 2495  %b29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 29), align 2496  %b30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 30), align 2497  %b31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 31), align 2498  %c0  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 0 ), align 2499  %c1  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 1 ), align 2500  %c2  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 2 ), align 2501  %c3  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 3 ), align 2502  %c4  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 4 ), align 2503  %c5  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 5 ), align 2504  %c6  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 6 ), align 2505  %c7  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 7 ), align 2506  %c8  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8 ), align 2507  %c9  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 9 ), align 2508  %c10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 10), align 2509  %c11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 11), align 2510  %c12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 12), align 2511  %c13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 13), align 2512  %c14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 14), align 2513  %c15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 15), align 2514  %c16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2515  %c17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 17), align 2516  %c18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 18), align 2517  %c19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 19), align 2518  %c20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 20), align 2519  %c21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 21), align 2520  %c22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 22), align 2521  %c23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 23), align 2522  %c24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2523  %c25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 25), align 2524  %c26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 26), align 2525  %c27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 27), align 2526  %c28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 28), align 2527  %c29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 29), align 2528  %c30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 30), align 2529  %c31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 31), align 2530  %r0  = call i16 @llvm.fshr.i16(i16 %a0 , i16 %b0 , i16 %c0 )531  %r1  = call i16 @llvm.fshr.i16(i16 %a1 , i16 %b1 , i16 %c1 )532  %r2  = call i16 @llvm.fshr.i16(i16 %a2 , i16 %b2 , i16 %c2 )533  %r3  = call i16 @llvm.fshr.i16(i16 %a3 , i16 %b3 , i16 %c3 )534  %r4  = call i16 @llvm.fshr.i16(i16 %a4 , i16 %b4 , i16 %c4 )535  %r5  = call i16 @llvm.fshr.i16(i16 %a5 , i16 %b5 , i16 %c5 )536  %r6  = call i16 @llvm.fshr.i16(i16 %a6 , i16 %b6 , i16 %c6 )537  %r7  = call i16 @llvm.fshr.i16(i16 %a7 , i16 %b7 , i16 %c7 )538  %r8  = call i16 @llvm.fshr.i16(i16 %a8 , i16 %b8 , i16 %c8 )539  %r9  = call i16 @llvm.fshr.i16(i16 %a9 , i16 %b9 , i16 %c9 )540  %r10 = call i16 @llvm.fshr.i16(i16 %a10, i16 %b10, i16 %c10)541  %r11 = call i16 @llvm.fshr.i16(i16 %a11, i16 %b11, i16 %c11)542  %r12 = call i16 @llvm.fshr.i16(i16 %a12, i16 %b12, i16 %c12)543  %r13 = call i16 @llvm.fshr.i16(i16 %a13, i16 %b13, i16 %c13)544  %r14 = call i16 @llvm.fshr.i16(i16 %a14, i16 %b14, i16 %c14)545  %r15 = call i16 @llvm.fshr.i16(i16 %a15, i16 %b15, i16 %c15)546  %r16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 %c16)547  %r17 = call i16 @llvm.fshr.i16(i16 %a17, i16 %b17, i16 %c17)548  %r18 = call i16 @llvm.fshr.i16(i16 %a18, i16 %b18, i16 %c18)549  %r19 = call i16 @llvm.fshr.i16(i16 %a19, i16 %b19, i16 %c19)550  %r20 = call i16 @llvm.fshr.i16(i16 %a20, i16 %b20, i16 %c20)551  %r21 = call i16 @llvm.fshr.i16(i16 %a21, i16 %b21, i16 %c21)552  %r22 = call i16 @llvm.fshr.i16(i16 %a22, i16 %b22, i16 %c22)553  %r23 = call i16 @llvm.fshr.i16(i16 %a23, i16 %b23, i16 %c23)554  %r24 = call i16 @llvm.fshr.i16(i16 %a24, i16 %b24, i16 %c24)555  %r25 = call i16 @llvm.fshr.i16(i16 %a25, i16 %b25, i16 %c25)556  %r26 = call i16 @llvm.fshr.i16(i16 %a26, i16 %b26, i16 %c26)557  %r27 = call i16 @llvm.fshr.i16(i16 %a27, i16 %b27, i16 %c27)558  %r28 = call i16 @llvm.fshr.i16(i16 %a28, i16 %b28, i16 %c28)559  %r29 = call i16 @llvm.fshr.i16(i16 %a29, i16 %b29, i16 %c29)560  %r30 = call i16 @llvm.fshr.i16(i16 %a30, i16 %b30, i16 %c30)561  %r31 = call i16 @llvm.fshr.i16(i16 %a31, i16 %b31, i16 %c31)562  store i16 %r0 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 0 ), align 2563  store i16 %r1 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 1 ), align 2564  store i16 %r2 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 2 ), align 2565  store i16 %r3 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 3 ), align 2566  store i16 %r4 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 4 ), align 2567  store i16 %r5 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 5 ), align 2568  store i16 %r6 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 6 ), align 2569  store i16 %r7 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 7 ), align 2570  store i16 %r8 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 8 ), align 2571  store i16 %r9 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 9 ), align 2572  store i16 %r10, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 10), align 2573  store i16 %r11, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 11), align 2574  store i16 %r12, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 12), align 2575  store i16 %r13, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 13), align 2576  store i16 %r14, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 14), align 2577  store i16 %r15, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 15), align 2578  store i16 %r16, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 16), align 2579  store i16 %r17, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 17), align 2580  store i16 %r18, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 18), align 2581  store i16 %r19, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 19), align 2582  store i16 %r20, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 20), align 2583  store i16 %r21, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 21), align 2584  store i16 %r22, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 22), align 2585  store i16 %r23, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 23), align 2586  store i16 %r24, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 24), align 2587  store i16 %r25, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 25), align 2588  store i16 %r26, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 26), align 2589  store i16 %r27, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 27), align 2590  store i16 %r28, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 28), align 2591  store i16 %r29, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 29), align 2592  store i16 %r30, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 30), align 2593  store i16 %r31, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 31), align 2594  ret void595}596 597define void @fshr_v64i8() {598; SSE-LABEL: @fshr_v64i8(599; SSE-NEXT:    [[TMP1:%.*]] = load <16 x i8>, ptr @a8, align 1600; SSE-NEXT:    [[TMP2:%.*]] = load <16 x i8>, ptr @b8, align 1601; SSE-NEXT:    [[TMP3:%.*]] = load <16 x i8>, ptr @c8, align 1602; SSE-NEXT:    [[TMP4:%.*]] = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> [[TMP1]], <16 x i8> [[TMP2]], <16 x i8> [[TMP3]])603; SSE-NEXT:    store <16 x i8> [[TMP4]], ptr @d8, align 1604; SSE-NEXT:    [[TMP5:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1605; SSE-NEXT:    [[TMP6:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1606; SSE-NEXT:    [[TMP7:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1607; SSE-NEXT:    [[TMP8:%.*]] = call <16 x i8> @llvm.fshr.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 1609; SSE-NEXT:    [[TMP9:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1610; SSE-NEXT:    [[TMP10:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1611; SSE-NEXT:    [[TMP11:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1612; SSE-NEXT:    [[TMP12:%.*]] = call <16 x i8> @llvm.fshr.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 1614; SSE-NEXT:    [[TMP13:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1615; SSE-NEXT:    [[TMP14:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1616; SSE-NEXT:    [[TMP15:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1617; SSE-NEXT:    [[TMP16:%.*]] = call <16 x i8> @llvm.fshr.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 1619; SSE-NEXT:    ret void620;621; AVX-LABEL: @fshr_v64i8(622; AVX-NEXT:    [[TMP1:%.*]] = load <32 x i8>, ptr @a8, align 1623; AVX-NEXT:    [[TMP2:%.*]] = load <32 x i8>, ptr @b8, align 1624; AVX-NEXT:    [[TMP3:%.*]] = load <32 x i8>, ptr @c8, align 1625; AVX-NEXT:    [[TMP4:%.*]] = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> [[TMP1]], <32 x i8> [[TMP2]], <32 x i8> [[TMP3]])626; AVX-NEXT:    store <32 x i8> [[TMP4]], ptr @d8, align 1627; AVX-NEXT:    [[TMP5:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1628; AVX-NEXT:    [[TMP6:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1629; AVX-NEXT:    [[TMP7:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1630; AVX-NEXT:    [[TMP8:%.*]] = call <32 x i8> @llvm.fshr.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 1632; AVX-NEXT:    ret void633;634; AVX512-LABEL: @fshr_v64i8(635; AVX512-NEXT:    [[TMP1:%.*]] = load <64 x i8>, ptr @a8, align 1636; AVX512-NEXT:    [[TMP2:%.*]] = load <64 x i8>, ptr @b8, align 1637; AVX512-NEXT:    [[TMP3:%.*]] = load <64 x i8>, ptr @c8, align 1638; AVX512-NEXT:    [[TMP4:%.*]] = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> [[TMP1]], <64 x i8> [[TMP2]], <64 x i8> [[TMP3]])639; AVX512-NEXT:    store <64 x i8> [[TMP4]], ptr @d8, align 1640; AVX512-NEXT:    ret void641;642; AVX512VBMI2-LABEL: @fshr_v64i8(643; AVX512VBMI2-NEXT:    [[TMP1:%.*]] = load <64 x i8>, ptr @a8, align 1644; AVX512VBMI2-NEXT:    [[TMP2:%.*]] = load <64 x i8>, ptr @b8, align 1645; AVX512VBMI2-NEXT:    [[TMP3:%.*]] = load <64 x i8>, ptr @c8, align 1646; AVX512VBMI2-NEXT:    [[TMP4:%.*]] = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> [[TMP1]], <64 x i8> [[TMP2]], <64 x i8> [[TMP3]])647; AVX512VBMI2-NEXT:    store <64 x i8> [[TMP4]], ptr @d8, align 1648; AVX512VBMI2-NEXT:    ret void649;650  %a0  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 0 ), align 1651  %a1  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 1 ), align 1652  %a2  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 2 ), align 1653  %a3  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 3 ), align 1654  %a4  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 4 ), align 1655  %a5  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 5 ), align 1656  %a6  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 6 ), align 1657  %a7  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 7 ), align 1658  %a8  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 8 ), align 1659  %a9  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 9 ), align 1660  %a10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 10), align 1661  %a11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 11), align 1662  %a12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 12), align 1663  %a13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 13), align 1664  %a14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 14), align 1665  %a15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 15), align 1666  %a16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1667  %a17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 17), align 1668  %a18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 18), align 1669  %a19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 19), align 1670  %a20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 20), align 1671  %a21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 21), align 1672  %a22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 22), align 1673  %a23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 23), align 1674  %a24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 24), align 1675  %a25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 25), align 1676  %a26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 26), align 1677  %a27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 27), align 1678  %a28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 28), align 1679  %a29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 29), align 1680  %a30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 30), align 1681  %a31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 31), align 1682  %a32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1683  %a33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 33), align 1684  %a34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 34), align 1685  %a35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 35), align 1686  %a36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 36), align 1687  %a37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 37), align 1688  %a38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 38), align 1689  %a39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 39), align 1690  %a40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 40), align 1691  %a41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 41), align 1692  %a42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 42), align 1693  %a43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 43), align 1694  %a44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 44), align 1695  %a45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 45), align 1696  %a46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 46), align 1697  %a47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 47), align 1698  %a48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1699  %a49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 49), align 1700  %a50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 50), align 1701  %a51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 51), align 1702  %a52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 52), align 1703  %a53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 53), align 1704  %a54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 54), align 1705  %a55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 55), align 1706  %a56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 56), align 1707  %a57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 57), align 1708  %a58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 58), align 1709  %a59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 59), align 1710  %a60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 60), align 1711  %a61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 61), align 1712  %a62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 62), align 1713  %a63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 63), align 1714  %b0  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 0 ), align 1715  %b1  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 1 ), align 1716  %b2  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 2 ), align 1717  %b3  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 3 ), align 1718  %b4  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 4 ), align 1719  %b5  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 5 ), align 1720  %b6  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 6 ), align 1721  %b7  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 7 ), align 1722  %b8  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 8 ), align 1723  %b9  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 9 ), align 1724  %b10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 10), align 1725  %b11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 11), align 1726  %b12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 12), align 1727  %b13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 13), align 1728  %b14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 14), align 1729  %b15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 15), align 1730  %b16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1731  %b17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 17), align 1732  %b18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 18), align 1733  %b19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 19), align 1734  %b20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 20), align 1735  %b21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 21), align 1736  %b22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 22), align 1737  %b23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 23), align 1738  %b24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 24), align 1739  %b25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 25), align 1740  %b26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 26), align 1741  %b27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 27), align 1742  %b28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 28), align 1743  %b29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 29), align 1744  %b30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 30), align 1745  %b31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 31), align 1746  %b32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1747  %b33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 33), align 1748  %b34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 34), align 1749  %b35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 35), align 1750  %b36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 36), align 1751  %b37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 37), align 1752  %b38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 38), align 1753  %b39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 39), align 1754  %b40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 40), align 1755  %b41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 41), align 1756  %b42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 42), align 1757  %b43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 43), align 1758  %b44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 44), align 1759  %b45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 45), align 1760  %b46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 46), align 1761  %b47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 47), align 1762  %b48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1763  %b49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 49), align 1764  %b50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 50), align 1765  %b51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 51), align 1766  %b52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 52), align 1767  %b53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 53), align 1768  %b54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 54), align 1769  %b55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 55), align 1770  %b56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 56), align 1771  %b57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 57), align 1772  %b58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 58), align 1773  %b59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 59), align 1774  %b60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 60), align 1775  %b61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 61), align 1776  %b62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 62), align 1777  %b63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 63), align 1778  %c0  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 0 ), align 1779  %c1  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 1 ), align 1780  %c2  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 2 ), align 1781  %c3  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 3 ), align 1782  %c4  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 4 ), align 1783  %c5  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 5 ), align 1784  %c6  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 6 ), align 1785  %c7  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 7 ), align 1786  %c8  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 8 ), align 1787  %c9  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 9 ), align 1788  %c10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 10), align 1789  %c11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 11), align 1790  %c12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 12), align 1791  %c13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 13), align 1792  %c14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 14), align 1793  %c15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 15), align 1794  %c16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1795  %c17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 17), align 1796  %c18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 18), align 1797  %c19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 19), align 1798  %c20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 20), align 1799  %c21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 21), align 1800  %c22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 22), align 1801  %c23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 23), align 1802  %c24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 24), align 1803  %c25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 25), align 1804  %c26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 26), align 1805  %c27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 27), align 1806  %c28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 28), align 1807  %c29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 29), align 1808  %c30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 30), align 1809  %c31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 31), align 1810  %c32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1811  %c33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 33), align 1812  %c34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 34), align 1813  %c35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 35), align 1814  %c36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 36), align 1815  %c37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 37), align 1816  %c38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 38), align 1817  %c39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 39), align 1818  %c40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 40), align 1819  %c41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 41), align 1820  %c42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 42), align 1821  %c43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 43), align 1822  %c44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 44), align 1823  %c45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 45), align 1824  %c46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 46), align 1825  %c47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 47), align 1826  %c48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1827  %c49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 49), align 1828  %c50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 50), align 1829  %c51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 51), align 1830  %c52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 52), align 1831  %c53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 53), align 1832  %c54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 54), align 1833  %c55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 55), align 1834  %c56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 56), align 1835  %c57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 57), align 1836  %c58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 58), align 1837  %c59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 59), align 1838  %c60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 60), align 1839  %c61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 61), align 1840  %c62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 62), align 1841  %c63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 63), align 1842  %r0  = call i8 @llvm.fshr.i8(i8 %a0 , i8 %b0 , i8 %c0 )843  %r1  = call i8 @llvm.fshr.i8(i8 %a1 , i8 %b1 , i8 %c1 )844  %r2  = call i8 @llvm.fshr.i8(i8 %a2 , i8 %b2 , i8 %c2 )845  %r3  = call i8 @llvm.fshr.i8(i8 %a3 , i8 %b3 , i8 %c3 )846  %r4  = call i8 @llvm.fshr.i8(i8 %a4 , i8 %b4 , i8 %c4 )847  %r5  = call i8 @llvm.fshr.i8(i8 %a5 , i8 %b5 , i8 %c5 )848  %r6  = call i8 @llvm.fshr.i8(i8 %a6 , i8 %b6 , i8 %c6 )849  %r7  = call i8 @llvm.fshr.i8(i8 %a7 , i8 %b7 , i8 %c7 )850  %r8  = call i8 @llvm.fshr.i8(i8 %a8 , i8 %b8 , i8 %c8 )851  %r9  = call i8 @llvm.fshr.i8(i8 %a9 , i8 %b9 , i8 %c9 )852  %r10 = call i8 @llvm.fshr.i8(i8 %a10, i8 %b10, i8 %c10)853  %r11 = call i8 @llvm.fshr.i8(i8 %a11, i8 %b11, i8 %c11)854  %r12 = call i8 @llvm.fshr.i8(i8 %a12, i8 %b12, i8 %c12)855  %r13 = call i8 @llvm.fshr.i8(i8 %a13, i8 %b13, i8 %c13)856  %r14 = call i8 @llvm.fshr.i8(i8 %a14, i8 %b14, i8 %c14)857  %r15 = call i8 @llvm.fshr.i8(i8 %a15, i8 %b15, i8 %c15)858  %r16 = call i8 @llvm.fshr.i8(i8 %a16, i8 %b16, i8 %c16)859  %r17 = call i8 @llvm.fshr.i8(i8 %a17, i8 %b17, i8 %c17)860  %r18 = call i8 @llvm.fshr.i8(i8 %a18, i8 %b18, i8 %c18)861  %r19 = call i8 @llvm.fshr.i8(i8 %a19, i8 %b19, i8 %c19)862  %r20 = call i8 @llvm.fshr.i8(i8 %a20, i8 %b20, i8 %c20)863  %r21 = call i8 @llvm.fshr.i8(i8 %a21, i8 %b21, i8 %c21)864  %r22 = call i8 @llvm.fshr.i8(i8 %a22, i8 %b22, i8 %c22)865  %r23 = call i8 @llvm.fshr.i8(i8 %a23, i8 %b23, i8 %c23)866  %r24 = call i8 @llvm.fshr.i8(i8 %a24, i8 %b24, i8 %c24)867  %r25 = call i8 @llvm.fshr.i8(i8 %a25, i8 %b25, i8 %c25)868  %r26 = call i8 @llvm.fshr.i8(i8 %a26, i8 %b26, i8 %c26)869  %r27 = call i8 @llvm.fshr.i8(i8 %a27, i8 %b27, i8 %c27)870  %r28 = call i8 @llvm.fshr.i8(i8 %a28, i8 %b28, i8 %c28)871  %r29 = call i8 @llvm.fshr.i8(i8 %a29, i8 %b29, i8 %c29)872  %r30 = call i8 @llvm.fshr.i8(i8 %a30, i8 %b30, i8 %c30)873  %r31 = call i8 @llvm.fshr.i8(i8 %a31, i8 %b31, i8 %c31)874  %r32 = call i8 @llvm.fshr.i8(i8 %a32, i8 %b32, i8 %c32)875  %r33 = call i8 @llvm.fshr.i8(i8 %a33, i8 %b33, i8 %c33)876  %r34 = call i8 @llvm.fshr.i8(i8 %a34, i8 %b34, i8 %c34)877  %r35 = call i8 @llvm.fshr.i8(i8 %a35, i8 %b35, i8 %c35)878  %r36 = call i8 @llvm.fshr.i8(i8 %a36, i8 %b36, i8 %c36)879  %r37 = call i8 @llvm.fshr.i8(i8 %a37, i8 %b37, i8 %c37)880  %r38 = call i8 @llvm.fshr.i8(i8 %a38, i8 %b38, i8 %c38)881  %r39 = call i8 @llvm.fshr.i8(i8 %a39, i8 %b39, i8 %c39)882  %r40 = call i8 @llvm.fshr.i8(i8 %a40, i8 %b40, i8 %c40)883  %r41 = call i8 @llvm.fshr.i8(i8 %a41, i8 %b41, i8 %c41)884  %r42 = call i8 @llvm.fshr.i8(i8 %a42, i8 %b42, i8 %c42)885  %r43 = call i8 @llvm.fshr.i8(i8 %a43, i8 %b43, i8 %c43)886  %r44 = call i8 @llvm.fshr.i8(i8 %a44, i8 %b44, i8 %c44)887  %r45 = call i8 @llvm.fshr.i8(i8 %a45, i8 %b45, i8 %c45)888  %r46 = call i8 @llvm.fshr.i8(i8 %a46, i8 %b46, i8 %c46)889  %r47 = call i8 @llvm.fshr.i8(i8 %a47, i8 %b47, i8 %c47)890  %r48 = call i8 @llvm.fshr.i8(i8 %a48, i8 %b48, i8 %c48)891  %r49 = call i8 @llvm.fshr.i8(i8 %a49, i8 %b49, i8 %c49)892  %r50 = call i8 @llvm.fshr.i8(i8 %a50, i8 %b50, i8 %c50)893  %r51 = call i8 @llvm.fshr.i8(i8 %a51, i8 %b51, i8 %c51)894  %r52 = call i8 @llvm.fshr.i8(i8 %a52, i8 %b52, i8 %c52)895  %r53 = call i8 @llvm.fshr.i8(i8 %a53, i8 %b53, i8 %c53)896  %r54 = call i8 @llvm.fshr.i8(i8 %a54, i8 %b54, i8 %c54)897  %r55 = call i8 @llvm.fshr.i8(i8 %a55, i8 %b55, i8 %c55)898  %r56 = call i8 @llvm.fshr.i8(i8 %a56, i8 %b56, i8 %c56)899  %r57 = call i8 @llvm.fshr.i8(i8 %a57, i8 %b57, i8 %c57)900  %r58 = call i8 @llvm.fshr.i8(i8 %a58, i8 %b58, i8 %c58)901  %r59 = call i8 @llvm.fshr.i8(i8 %a59, i8 %b59, i8 %c59)902  %r60 = call i8 @llvm.fshr.i8(i8 %a60, i8 %b60, i8 %c60)903  %r61 = call i8 @llvm.fshr.i8(i8 %a61, i8 %b61, i8 %c61)904  %r62 = call i8 @llvm.fshr.i8(i8 %a62, i8 %b62, i8 %c62)905  %r63 = call i8 @llvm.fshr.i8(i8 %a63, i8 %b63, i8 %c63)906  store i8 %r0 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 0 ), align 1907  store i8 %r1 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 1 ), align 1908  store i8 %r2 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 2 ), align 1909  store i8 %r3 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 3 ), align 1910  store i8 %r4 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 4 ), align 1911  store i8 %r5 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 5 ), align 1912  store i8 %r6 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 6 ), align 1913  store i8 %r7 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 7 ), align 1914  store i8 %r8 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 8 ), align 1915  store i8 %r9 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 9 ), align 1916  store i8 %r10, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 10), align 1917  store i8 %r11, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 11), align 1918  store i8 %r12, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 12), align 1919  store i8 %r13, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 13), align 1920  store i8 %r14, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 14), align 1921  store i8 %r15, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 15), align 1922  store i8 %r16, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 16), align 1923  store i8 %r17, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 17), align 1924  store i8 %r18, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 18), align 1925  store i8 %r19, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 19), align 1926  store i8 %r20, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 20), align 1927  store i8 %r21, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 21), align 1928  store i8 %r22, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 22), align 1929  store i8 %r23, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 23), align 1930  store i8 %r24, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 24), align 1931  store i8 %r25, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 25), align 1932  store i8 %r26, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 26), align 1933  store i8 %r27, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 27), align 1934  store i8 %r28, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 28), align 1935  store i8 %r29, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 29), align 1936  store i8 %r30, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 30), align 1937  store i8 %r31, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 31), align 1938  store i8 %r32, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 32), align 1939  store i8 %r33, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 33), align 1940  store i8 %r34, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 34), align 1941  store i8 %r35, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 35), align 1942  store i8 %r36, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 36), align 1943  store i8 %r37, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 37), align 1944  store i8 %r38, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 38), align 1945  store i8 %r39, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 39), align 1946  store i8 %r40, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 40), align 1947  store i8 %r41, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 41), align 1948  store i8 %r42, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 42), align 1949  store i8 %r43, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 43), align 1950  store i8 %r44, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 44), align 1951  store i8 %r45, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 45), align 1952  store i8 %r46, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 46), align 1953  store i8 %r47, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 47), align 1954  store i8 %r48, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 48), align 1955  store i8 %r49, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 49), align 1956  store i8 %r50, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 50), align 1957  store i8 %r51, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 51), align 1958  store i8 %r52, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 52), align 1959  store i8 %r53, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 53), align 1960  store i8 %r54, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 54), align 1961  store i8 %r55, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 55), align 1962  store i8 %r56, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 56), align 1963  store i8 %r57, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 57), align 1964  store i8 %r58, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 58), align 1965  store i8 %r59, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 59), align 1966  store i8 %r60, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 60), align 1967  store i8 %r61, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 61), align 1968  store i8 %r62, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 62), align 1969  store i8 %r63, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 63), align 1970  ret void971}972