brintos / llvm-project-archived public Read only

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=SLM4; 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=knl -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX5127; 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=skx -mattr=+prefer-256-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX,AVX256BW9 10@a64 = common global [8 x i64] zeroinitializer, align 6411@b64 = common global [8 x i64] zeroinitializer, align 6412@c64 = common global [8 x i64] zeroinitializer, align 6413@a32 = common global [16 x i32] zeroinitializer, align 6414@b32 = common global [16 x i32] zeroinitializer, align 6415@c32 = common global [16 x i32] zeroinitializer, align 6416@a16 = common global [32 x i16] zeroinitializer, align 6417@b16 = common global [32 x i16] zeroinitializer, align 6418@c16 = common global [32 x i16] zeroinitializer, align 6419@a8  = common global [64 x i8] zeroinitializer, align 6420@b8  = common global [64 x i8] zeroinitializer, align 6421@c8  = common global [64 x i8] zeroinitializer, align 6422 23declare i64 @llvm.smul.fix.i64(i64, i64, i32)24declare i32 @llvm.smul.fix.i32(i32, i32, i32)25declare i16 @llvm.smul.fix.i16(i16, i16, i32)26declare i8  @llvm.smul.fix.i8 (i8 , i8 , i32)27 28define void @smul_v8i64() {29; SSE-LABEL: @smul_v8i64(30; SSE-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @a64, align 831; SSE-NEXT:    [[TMP2:%.*]] = load <2 x i64>, ptr @b64, align 832; SSE-NEXT:    [[TMP3:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP1]], <2 x i64> [[TMP2]], i32 3)33; SSE-NEXT:    store <2 x i64> [[TMP3]], ptr @c64, align 834; SSE-NEXT:    [[TMP4:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 835; SSE-NEXT:    [[TMP5:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 836; SSE-NEXT:    [[TMP6:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP4]], <2 x i64> [[TMP5]], i32 3)37; SSE-NEXT:    store <2 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 838; SSE-NEXT:    [[TMP7:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 839; SSE-NEXT:    [[TMP8:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 840; SSE-NEXT:    [[TMP9:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP7]], <2 x i64> [[TMP8]], i32 3)41; SSE-NEXT:    store <2 x i64> [[TMP9]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 842; SSE-NEXT:    [[TMP10:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 843; SSE-NEXT:    [[TMP11:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 844; SSE-NEXT:    [[TMP12:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP10]], <2 x i64> [[TMP11]], i32 3)45; SSE-NEXT:    store <2 x i64> [[TMP12]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 846; SSE-NEXT:    ret void47;48; SLM-LABEL: @smul_v8i64(49; SLM-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @a64, align 850; SLM-NEXT:    [[TMP2:%.*]] = load <2 x i64>, ptr @b64, align 851; SLM-NEXT:    [[TMP3:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP1]], <2 x i64> [[TMP2]], i32 3)52; SLM-NEXT:    store <2 x i64> [[TMP3]], ptr @c64, align 853; SLM-NEXT:    [[TMP4:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 854; SLM-NEXT:    [[TMP5:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 855; SLM-NEXT:    [[TMP6:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP4]], <2 x i64> [[TMP5]], i32 3)56; SLM-NEXT:    store <2 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 857; SLM-NEXT:    [[TMP7:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 858; SLM-NEXT:    [[TMP8:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 859; SLM-NEXT:    [[TMP9:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP7]], <2 x i64> [[TMP8]], i32 3)60; SLM-NEXT:    store <2 x i64> [[TMP9]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 861; SLM-NEXT:    [[TMP10:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 862; SLM-NEXT:    [[TMP11:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 863; SLM-NEXT:    [[TMP12:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP10]], <2 x i64> [[TMP11]], i32 3)64; SLM-NEXT:    store <2 x i64> [[TMP12]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 865; SLM-NEXT:    ret void66;67; AVX1-LABEL: @smul_v8i64(68; AVX1-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @a64, align 869; AVX1-NEXT:    [[TMP2:%.*]] = load <2 x i64>, ptr @b64, align 870; AVX1-NEXT:    [[TMP3:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP1]], <2 x i64> [[TMP2]], i32 3)71; AVX1-NEXT:    store <2 x i64> [[TMP3]], ptr @c64, align 872; AVX1-NEXT:    [[TMP4:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 873; AVX1-NEXT:    [[TMP5:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 874; AVX1-NEXT:    [[TMP6:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP4]], <2 x i64> [[TMP5]], i32 3)75; AVX1-NEXT:    store <2 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 876; AVX1-NEXT:    [[TMP7:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 877; AVX1-NEXT:    [[TMP8:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 878; AVX1-NEXT:    [[TMP9:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP7]], <2 x i64> [[TMP8]], i32 3)79; AVX1-NEXT:    store <2 x i64> [[TMP9]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 880; AVX1-NEXT:    [[TMP10:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 881; AVX1-NEXT:    [[TMP11:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 882; AVX1-NEXT:    [[TMP12:%.*]] = call <2 x i64> @llvm.smul.fix.v2i64(<2 x i64> [[TMP10]], <2 x i64> [[TMP11]], i32 3)83; AVX1-NEXT:    store <2 x i64> [[TMP12]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 884; AVX1-NEXT:    ret void85;86; AVX2-LABEL: @smul_v8i64(87; AVX2-NEXT:    [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 888; AVX2-NEXT:    [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 889; AVX2-NEXT:    [[TMP3:%.*]] = call <4 x i64> @llvm.smul.fix.v4i64(<4 x i64> [[TMP1]], <4 x i64> [[TMP2]], i32 3)90; AVX2-NEXT:    store <4 x i64> [[TMP3]], ptr @c64, align 891; AVX2-NEXT:    [[TMP4:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 892; AVX2-NEXT:    [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 893; AVX2-NEXT:    [[TMP6:%.*]] = call <4 x i64> @llvm.smul.fix.v4i64(<4 x i64> [[TMP4]], <4 x i64> [[TMP5]], i32 3)94; AVX2-NEXT:    store <4 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 895; AVX2-NEXT:    ret void96;97; AVX512-LABEL: @smul_v8i64(98; AVX512-NEXT:    [[TMP1:%.*]] = load <8 x i64>, ptr @a64, align 899; AVX512-NEXT:    [[TMP2:%.*]] = load <8 x i64>, ptr @b64, align 8100; AVX512-NEXT:    [[TMP3:%.*]] = call <8 x i64> @llvm.smul.fix.v8i64(<8 x i64> [[TMP1]], <8 x i64> [[TMP2]], i32 3)101; AVX512-NEXT:    store <8 x i64> [[TMP3]], ptr @c64, align 8102; AVX512-NEXT:    ret void103;104; AVX256BW-LABEL: @smul_v8i64(105; AVX256BW-NEXT:    [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 8106; AVX256BW-NEXT:    [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 8107; AVX256BW-NEXT:    [[TMP3:%.*]] = call <4 x i64> @llvm.smul.fix.v4i64(<4 x i64> [[TMP1]], <4 x i64> [[TMP2]], i32 3)108; AVX256BW-NEXT:    store <4 x i64> [[TMP3]], ptr @c64, align 8109; AVX256BW-NEXT:    [[TMP4:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8110; AVX256BW-NEXT:    [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8111; AVX256BW-NEXT:    [[TMP6:%.*]] = call <4 x i64> @llvm.smul.fix.v4i64(<4 x i64> [[TMP4]], <4 x i64> [[TMP5]], i32 3)112; AVX256BW-NEXT:    store <4 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8113; AVX256BW-NEXT:    ret void114;115  %a0 = load i64, ptr @a64, align 8116  %a1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 8117  %a2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8118  %a3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 8119  %a4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8120  %a5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 8121  %a6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8122  %a7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 8123  %b0 = load i64, ptr @b64, align 8124  %b1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 8125  %b2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8126  %b3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 8127  %b4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8128  %b5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 8129  %b6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8130  %b7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 8131  %r0 = call i64 @llvm.smul.fix.i64(i64 %a0, i64 %b0, i32 3)132  %r1 = call i64 @llvm.smul.fix.i64(i64 %a1, i64 %b1, i32 3)133  %r2 = call i64 @llvm.smul.fix.i64(i64 %a2, i64 %b2, i32 3)134  %r3 = call i64 @llvm.smul.fix.i64(i64 %a3, i64 %b3, i32 3)135  %r4 = call i64 @llvm.smul.fix.i64(i64 %a4, i64 %b4, i32 3)136  %r5 = call i64 @llvm.smul.fix.i64(i64 %a5, i64 %b5, i32 3)137  %r6 = call i64 @llvm.smul.fix.i64(i64 %a6, i64 %b6, i32 3)138  %r7 = call i64 @llvm.smul.fix.i64(i64 %a7, i64 %b7, i32 3)139  store i64 %r0, ptr @c64, align 8140  store i64 %r1, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 1), align 8141  store i64 %r2, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8142  store i64 %r3, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 3), align 8143  store i64 %r4, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8144  store i64 %r5, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 5), align 8145  store i64 %r6, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8146  store i64 %r7, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 7), align 8147  ret void148}149 150define void @smul_v16i32() {151; SSE-LABEL: @smul_v16i32(152; SSE-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr @a32, align 4153; SSE-NEXT:    [[TMP2:%.*]] = load <4 x i32>, ptr @b32, align 4154; SSE-NEXT:    [[TMP3:%.*]] = call <4 x i32> @llvm.smul.fix.v4i32(<4 x i32> [[TMP1]], <4 x i32> [[TMP2]], i32 3)155; SSE-NEXT:    store <4 x i32> [[TMP3]], ptr @c32, align 4156; SSE-NEXT:    [[TMP4:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4157; SSE-NEXT:    [[TMP5:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4158; SSE-NEXT:    [[TMP6:%.*]] = call <4 x i32> @llvm.smul.fix.v4i32(<4 x i32> [[TMP4]], <4 x i32> [[TMP5]], i32 3)159; SSE-NEXT:    store <4 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4), align 4160; SSE-NEXT:    [[TMP7:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4161; SSE-NEXT:    [[TMP8:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4162; SSE-NEXT:    [[TMP9:%.*]] = call <4 x i32> @llvm.smul.fix.v4i32(<4 x i32> [[TMP7]], <4 x i32> [[TMP8]], i32 3)163; SSE-NEXT:    store <4 x i32> [[TMP9]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4164; SSE-NEXT:    [[TMP10:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4165; SSE-NEXT:    [[TMP11:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4166; SSE-NEXT:    [[TMP12:%.*]] = call <4 x i32> @llvm.smul.fix.v4i32(<4 x i32> [[TMP10]], <4 x i32> [[TMP11]], i32 3)167; SSE-NEXT:    store <4 x i32> [[TMP12]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4168; SSE-NEXT:    ret void169;170; SLM-LABEL: @smul_v16i32(171; SLM-NEXT:    [[A0:%.*]] = load i32, ptr @a32, align 4172; SLM-NEXT:    [[A1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1), align 4173; SLM-NEXT:    [[A2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2), align 4174; SLM-NEXT:    [[A3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3), align 4175; SLM-NEXT:    [[A4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4176; SLM-NEXT:    [[A5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5), align 4177; SLM-NEXT:    [[A6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6), align 4178; SLM-NEXT:    [[A7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7), align 4179; SLM-NEXT:    [[A8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4180; SLM-NEXT:    [[A9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9), align 4181; SLM-NEXT:    [[A10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 4182; SLM-NEXT:    [[A11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 4183; SLM-NEXT:    [[A12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4184; SLM-NEXT:    [[A13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 4185; SLM-NEXT:    [[A14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 4186; SLM-NEXT:    [[A15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 4187; SLM-NEXT:    [[B0:%.*]] = load i32, ptr @b32, align 4188; SLM-NEXT:    [[B1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1), align 4189; SLM-NEXT:    [[B2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2), align 4190; SLM-NEXT:    [[B3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3), align 4191; SLM-NEXT:    [[B4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4192; SLM-NEXT:    [[B5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5), align 4193; SLM-NEXT:    [[B6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6), align 4194; SLM-NEXT:    [[B7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7), align 4195; SLM-NEXT:    [[B8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4196; SLM-NEXT:    [[B9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9), align 4197; SLM-NEXT:    [[B10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 4198; SLM-NEXT:    [[B11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 4199; SLM-NEXT:    [[B12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4200; SLM-NEXT:    [[B13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 4201; SLM-NEXT:    [[B14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 4202; SLM-NEXT:    [[B15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 4203; SLM-NEXT:    [[R0:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A0]], i32 [[B0]], i32 3)204; SLM-NEXT:    [[R1:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A1]], i32 [[B1]], i32 3)205; SLM-NEXT:    [[R2:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A2]], i32 [[B2]], i32 3)206; SLM-NEXT:    [[R3:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A3]], i32 [[B3]], i32 3)207; SLM-NEXT:    [[R4:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A4]], i32 [[B4]], i32 3)208; SLM-NEXT:    [[R5:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A5]], i32 [[B5]], i32 3)209; SLM-NEXT:    [[R6:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A6]], i32 [[B6]], i32 3)210; SLM-NEXT:    [[R7:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A7]], i32 [[B7]], i32 3)211; SLM-NEXT:    [[R8:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A8]], i32 [[B8]], i32 3)212; SLM-NEXT:    [[R9:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A9]], i32 [[B9]], i32 3)213; SLM-NEXT:    [[R10:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A10]], i32 [[B10]], i32 3)214; SLM-NEXT:    [[R11:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A11]], i32 [[B11]], i32 3)215; SLM-NEXT:    [[R12:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A12]], i32 [[B12]], i32 3)216; SLM-NEXT:    [[R13:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A13]], i32 [[B13]], i32 3)217; SLM-NEXT:    [[R14:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A14]], i32 [[B14]], i32 3)218; SLM-NEXT:    [[R15:%.*]] = call i32 @llvm.smul.fix.i32(i32 [[A15]], i32 [[B15]], i32 3)219; SLM-NEXT:    store i32 [[R0]], ptr @c32, align 4220; SLM-NEXT:    store i32 [[R1]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 1), align 4221; SLM-NEXT:    store i32 [[R2]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 2), align 4222; SLM-NEXT:    store i32 [[R3]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 3), align 4223; SLM-NEXT:    store i32 [[R4]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4), align 4224; SLM-NEXT:    store i32 [[R5]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 5), align 4225; SLM-NEXT:    store i32 [[R6]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 6), align 4226; SLM-NEXT:    store i32 [[R7]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 7), align 4227; SLM-NEXT:    store i32 [[R8]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4228; SLM-NEXT:    store i32 [[R9]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 9), align 4229; SLM-NEXT:    store i32 [[R10]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 10), align 4230; SLM-NEXT:    store i32 [[R11]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 11), align 4231; SLM-NEXT:    store i32 [[R12]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4232; SLM-NEXT:    store i32 [[R13]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 13), align 4233; SLM-NEXT:    store i32 [[R14]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 14), align 4234; SLM-NEXT:    store i32 [[R15]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 15), align 4235; SLM-NEXT:    ret void236;237; AVX-LABEL: @smul_v16i32(238; AVX-NEXT:    [[TMP1:%.*]] = load <8 x i32>, ptr @a32, align 4239; AVX-NEXT:    [[TMP2:%.*]] = load <8 x i32>, ptr @b32, align 4240; AVX-NEXT:    [[TMP3:%.*]] = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> [[TMP1]], <8 x i32> [[TMP2]], i32 3)241; AVX-NEXT:    store <8 x i32> [[TMP3]], ptr @c32, align 4242; AVX-NEXT:    [[TMP4:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4243; AVX-NEXT:    [[TMP5:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4244; AVX-NEXT:    [[TMP6:%.*]] = call <8 x i32> @llvm.smul.fix.v8i32(<8 x i32> [[TMP4]], <8 x i32> [[TMP5]], i32 3)245; AVX-NEXT:    store <8 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4246; AVX-NEXT:    ret void247;248; AVX512-LABEL: @smul_v16i32(249; AVX512-NEXT:    [[TMP1:%.*]] = load <16 x i32>, ptr @a32, align 4250; AVX512-NEXT:    [[TMP2:%.*]] = load <16 x i32>, ptr @b32, align 4251; AVX512-NEXT:    [[TMP3:%.*]] = call <16 x i32> @llvm.smul.fix.v16i32(<16 x i32> [[TMP1]], <16 x i32> [[TMP2]], i32 3)252; AVX512-NEXT:    store <16 x i32> [[TMP3]], ptr @c32, align 4253; AVX512-NEXT:    ret void254;255  %a0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 0 ), align 4256  %a1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1 ), align 4257  %a2  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2 ), align 4258  %a3  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3 ), align 4259  %a4  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4 ), align 4260  %a5  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5 ), align 4261  %a6  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6 ), align 4262  %a7  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7 ), align 4263  %a8  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8 ), align 4264  %a9  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9 ), align 4265  %a10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 4266  %a11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 4267  %a12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4268  %a13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 4269  %a14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 4270  %a15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 4271  %b0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 0 ), align 4272  %b1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1 ), align 4273  %b2  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2 ), align 4274  %b3  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3 ), align 4275  %b4  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4 ), align 4276  %b5  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5 ), align 4277  %b6  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6 ), align 4278  %b7  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7 ), align 4279  %b8  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8 ), align 4280  %b9  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9 ), align 4281  %b10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 4282  %b11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 4283  %b12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4284  %b13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 4285  %b14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 4286  %b15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 4287  %r0  = call i32 @llvm.smul.fix.i32(i32 %a0 , i32 %b0 , i32 3)288  %r1  = call i32 @llvm.smul.fix.i32(i32 %a1 , i32 %b1 , i32 3)289  %r2  = call i32 @llvm.smul.fix.i32(i32 %a2 , i32 %b2 , i32 3)290  %r3  = call i32 @llvm.smul.fix.i32(i32 %a3 , i32 %b3 , i32 3)291  %r4  = call i32 @llvm.smul.fix.i32(i32 %a4 , i32 %b4 , i32 3)292  %r5  = call i32 @llvm.smul.fix.i32(i32 %a5 , i32 %b5 , i32 3)293  %r6  = call i32 @llvm.smul.fix.i32(i32 %a6 , i32 %b6 , i32 3)294  %r7  = call i32 @llvm.smul.fix.i32(i32 %a7 , i32 %b7 , i32 3)295  %r8  = call i32 @llvm.smul.fix.i32(i32 %a8 , i32 %b8 , i32 3)296  %r9  = call i32 @llvm.smul.fix.i32(i32 %a9 , i32 %b9 , i32 3)297  %r10 = call i32 @llvm.smul.fix.i32(i32 %a10, i32 %b10, i32 3)298  %r11 = call i32 @llvm.smul.fix.i32(i32 %a11, i32 %b11, i32 3)299  %r12 = call i32 @llvm.smul.fix.i32(i32 %a12, i32 %b12, i32 3)300  %r13 = call i32 @llvm.smul.fix.i32(i32 %a13, i32 %b13, i32 3)301  %r14 = call i32 @llvm.smul.fix.i32(i32 %a14, i32 %b14, i32 3)302  %r15 = call i32 @llvm.smul.fix.i32(i32 %a15, i32 %b15, i32 3)303  store i32 %r0 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 0 ), align 4304  store i32 %r1 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 1 ), align 4305  store i32 %r2 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 2 ), align 4306  store i32 %r3 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 3 ), align 4307  store i32 %r4 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4 ), align 4308  store i32 %r5 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 5 ), align 4309  store i32 %r6 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 6 ), align 4310  store i32 %r7 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 7 ), align 4311  store i32 %r8 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8 ), align 4312  store i32 %r9 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 9 ), align 4313  store i32 %r10, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 10), align 4314  store i32 %r11, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 11), align 4315  store i32 %r12, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4316  store i32 %r13, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 13), align 4317  store i32 %r14, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 14), align 4318  store i32 %r15, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 15), align 4319  ret void320}321 322define void @smul_v32i16() {323; SSE-LABEL: @smul_v32i16(324; SSE-NEXT:    [[TMP1:%.*]] = load <8 x i16>, ptr @a16, align 2325; SSE-NEXT:    [[TMP2:%.*]] = load <8 x i16>, ptr @b16, align 2326; SSE-NEXT:    [[TMP3:%.*]] = call <8 x i16> @llvm.smul.fix.v8i16(<8 x i16> [[TMP1]], <8 x i16> [[TMP2]], i32 3)327; SSE-NEXT:    store <8 x i16> [[TMP3]], ptr @c16, align 2328; SSE-NEXT:    [[TMP4:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8), align 2329; SSE-NEXT:    [[TMP5:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8), align 2330; SSE-NEXT:    [[TMP6:%.*]] = call <8 x i16> @llvm.smul.fix.v8i16(<8 x i16> [[TMP4]], <8 x i16> [[TMP5]], i32 3)331; SSE-NEXT:    store <8 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8), align 2332; SSE-NEXT:    [[TMP7:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2333; SSE-NEXT:    [[TMP8:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2334; SSE-NEXT:    [[TMP9:%.*]] = call <8 x i16> @llvm.smul.fix.v8i16(<8 x i16> [[TMP7]], <8 x i16> [[TMP8]], i32 3)335; SSE-NEXT:    store <8 x i16> [[TMP9]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2336; SSE-NEXT:    [[TMP10:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2337; SSE-NEXT:    [[TMP11:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2338; SSE-NEXT:    [[TMP12:%.*]] = call <8 x i16> @llvm.smul.fix.v8i16(<8 x i16> [[TMP10]], <8 x i16> [[TMP11]], i32 3)339; SSE-NEXT:    store <8 x i16> [[TMP12]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2340; SSE-NEXT:    ret void341;342; SLM-LABEL: @smul_v32i16(343; SLM-NEXT:    [[TMP1:%.*]] = load <8 x i16>, ptr @a16, align 2344; SLM-NEXT:    [[TMP2:%.*]] = load <8 x i16>, ptr @b16, align 2345; SLM-NEXT:    [[TMP3:%.*]] = call <8 x i16> @llvm.smul.fix.v8i16(<8 x i16> [[TMP1]], <8 x i16> [[TMP2]], i32 3)346; SLM-NEXT:    store <8 x i16> [[TMP3]], ptr @c16, align 2347; SLM-NEXT:    [[TMP4:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8), align 2348; SLM-NEXT:    [[TMP5:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8), align 2349; SLM-NEXT:    [[TMP6:%.*]] = call <8 x i16> @llvm.smul.fix.v8i16(<8 x i16> [[TMP4]], <8 x i16> [[TMP5]], i32 3)350; SLM-NEXT:    store <8 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8), align 2351; SLM-NEXT:    [[TMP7:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2352; SLM-NEXT:    [[TMP8:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2353; SLM-NEXT:    [[TMP9:%.*]] = call <8 x i16> @llvm.smul.fix.v8i16(<8 x i16> [[TMP7]], <8 x i16> [[TMP8]], i32 3)354; SLM-NEXT:    store <8 x i16> [[TMP9]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2355; SLM-NEXT:    [[TMP10:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2356; SLM-NEXT:    [[TMP11:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2357; SLM-NEXT:    [[TMP12:%.*]] = call <8 x i16> @llvm.smul.fix.v8i16(<8 x i16> [[TMP10]], <8 x i16> [[TMP11]], i32 3)358; SLM-NEXT:    store <8 x i16> [[TMP12]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2359; SLM-NEXT:    ret void360;361; AVX-LABEL: @smul_v32i16(362; AVX-NEXT:    [[TMP1:%.*]] = load <16 x i16>, ptr @a16, align 2363; AVX-NEXT:    [[TMP2:%.*]] = load <16 x i16>, ptr @b16, align 2364; AVX-NEXT:    [[TMP3:%.*]] = call <16 x i16> @llvm.smul.fix.v16i16(<16 x i16> [[TMP1]], <16 x i16> [[TMP2]], i32 3)365; AVX-NEXT:    store <16 x i16> [[TMP3]], ptr @c16, align 2366; AVX-NEXT:    [[TMP4:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2367; AVX-NEXT:    [[TMP5:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2368; AVX-NEXT:    [[TMP6:%.*]] = call <16 x i16> @llvm.smul.fix.v16i16(<16 x i16> [[TMP4]], <16 x i16> [[TMP5]], i32 3)369; AVX-NEXT:    store <16 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2370; AVX-NEXT:    ret void371;372; AVX512-LABEL: @smul_v32i16(373; AVX512-NEXT:    [[TMP1:%.*]] = load <32 x i16>, ptr @a16, align 2374; AVX512-NEXT:    [[TMP2:%.*]] = load <32 x i16>, ptr @b16, align 2375; AVX512-NEXT:    [[TMP3:%.*]] = call <32 x i16> @llvm.smul.fix.v32i16(<32 x i16> [[TMP1]], <32 x i16> [[TMP2]], i32 3)376; AVX512-NEXT:    store <32 x i16> [[TMP3]], ptr @c16, align 2377; AVX512-NEXT:    ret void378;379  %a0  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 0 ), align 2380  %a1  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 1 ), align 2381  %a2  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 2 ), align 2382  %a3  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 3 ), align 2383  %a4  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 4 ), align 2384  %a5  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 5 ), align 2385  %a6  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 6 ), align 2386  %a7  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 7 ), align 2387  %a8  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8 ), align 2388  %a9  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 9 ), align 2389  %a10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 10), align 2390  %a11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 11), align 2391  %a12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 12), align 2392  %a13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 13), align 2393  %a14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 14), align 2394  %a15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 15), align 2395  %a16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2396  %a17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 17), align 2397  %a18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 18), align 2398  %a19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 19), align 2399  %a20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 20), align 2400  %a21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 21), align 2401  %a22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 22), align 2402  %a23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 23), align 2403  %a24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2404  %a25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 25), align 2405  %a26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 26), align 2406  %a27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 27), align 2407  %a28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 28), align 2408  %a29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 29), align 2409  %a30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 30), align 2410  %a31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 31), align 2411  %b0  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 0 ), align 2412  %b1  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 1 ), align 2413  %b2  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 2 ), align 2414  %b3  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 3 ), align 2415  %b4  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 4 ), align 2416  %b5  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 5 ), align 2417  %b6  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 6 ), align 2418  %b7  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 7 ), align 2419  %b8  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8 ), align 2420  %b9  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 9 ), align 2421  %b10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 10), align 2422  %b11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 11), align 2423  %b12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 12), align 2424  %b13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 13), align 2425  %b14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 14), align 2426  %b15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 15), align 2427  %b16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2428  %b17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 17), align 2429  %b18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 18), align 2430  %b19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 19), align 2431  %b20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 20), align 2432  %b21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 21), align 2433  %b22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 22), align 2434  %b23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 23), align 2435  %b24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2436  %b25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 25), align 2437  %b26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 26), align 2438  %b27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 27), align 2439  %b28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 28), align 2440  %b29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 29), align 2441  %b30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 30), align 2442  %b31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 31), align 2443  %r0  = call i16 @llvm.smul.fix.i16(i16 %a0 , i16 %b0 , i32 3)444  %r1  = call i16 @llvm.smul.fix.i16(i16 %a1 , i16 %b1 , i32 3)445  %r2  = call i16 @llvm.smul.fix.i16(i16 %a2 , i16 %b2 , i32 3)446  %r3  = call i16 @llvm.smul.fix.i16(i16 %a3 , i16 %b3 , i32 3)447  %r4  = call i16 @llvm.smul.fix.i16(i16 %a4 , i16 %b4 , i32 3)448  %r5  = call i16 @llvm.smul.fix.i16(i16 %a5 , i16 %b5 , i32 3)449  %r6  = call i16 @llvm.smul.fix.i16(i16 %a6 , i16 %b6 , i32 3)450  %r7  = call i16 @llvm.smul.fix.i16(i16 %a7 , i16 %b7 , i32 3)451  %r8  = call i16 @llvm.smul.fix.i16(i16 %a8 , i16 %b8 , i32 3)452  %r9  = call i16 @llvm.smul.fix.i16(i16 %a9 , i16 %b9 , i32 3)453  %r10 = call i16 @llvm.smul.fix.i16(i16 %a10, i16 %b10, i32 3)454  %r11 = call i16 @llvm.smul.fix.i16(i16 %a11, i16 %b11, i32 3)455  %r12 = call i16 @llvm.smul.fix.i16(i16 %a12, i16 %b12, i32 3)456  %r13 = call i16 @llvm.smul.fix.i16(i16 %a13, i16 %b13, i32 3)457  %r14 = call i16 @llvm.smul.fix.i16(i16 %a14, i16 %b14, i32 3)458  %r15 = call i16 @llvm.smul.fix.i16(i16 %a15, i16 %b15, i32 3)459  %r16 = call i16 @llvm.smul.fix.i16(i16 %a16, i16 %b16, i32 3)460  %r17 = call i16 @llvm.smul.fix.i16(i16 %a17, i16 %b17, i32 3)461  %r18 = call i16 @llvm.smul.fix.i16(i16 %a18, i16 %b18, i32 3)462  %r19 = call i16 @llvm.smul.fix.i16(i16 %a19, i16 %b19, i32 3)463  %r20 = call i16 @llvm.smul.fix.i16(i16 %a20, i16 %b20, i32 3)464  %r21 = call i16 @llvm.smul.fix.i16(i16 %a21, i16 %b21, i32 3)465  %r22 = call i16 @llvm.smul.fix.i16(i16 %a22, i16 %b22, i32 3)466  %r23 = call i16 @llvm.smul.fix.i16(i16 %a23, i16 %b23, i32 3)467  %r24 = call i16 @llvm.smul.fix.i16(i16 %a24, i16 %b24, i32 3)468  %r25 = call i16 @llvm.smul.fix.i16(i16 %a25, i16 %b25, i32 3)469  %r26 = call i16 @llvm.smul.fix.i16(i16 %a26, i16 %b26, i32 3)470  %r27 = call i16 @llvm.smul.fix.i16(i16 %a27, i16 %b27, i32 3)471  %r28 = call i16 @llvm.smul.fix.i16(i16 %a28, i16 %b28, i32 3)472  %r29 = call i16 @llvm.smul.fix.i16(i16 %a29, i16 %b29, i32 3)473  %r30 = call i16 @llvm.smul.fix.i16(i16 %a30, i16 %b30, i32 3)474  %r31 = call i16 @llvm.smul.fix.i16(i16 %a31, i16 %b31, i32 3)475  store i16 %r0 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 0 ), align 2476  store i16 %r1 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 1 ), align 2477  store i16 %r2 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 2 ), align 2478  store i16 %r3 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 3 ), align 2479  store i16 %r4 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 4 ), align 2480  store i16 %r5 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 5 ), align 2481  store i16 %r6 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 6 ), align 2482  store i16 %r7 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 7 ), align 2483  store i16 %r8 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8 ), align 2484  store i16 %r9 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 9 ), align 2485  store i16 %r10, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 10), align 2486  store i16 %r11, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 11), align 2487  store i16 %r12, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 12), align 2488  store i16 %r13, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 13), align 2489  store i16 %r14, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 14), align 2490  store i16 %r15, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 15), align 2491  store i16 %r16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 2492  store i16 %r17, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 17), align 2493  store i16 %r18, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 18), align 2494  store i16 %r19, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 19), align 2495  store i16 %r20, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 20), align 2496  store i16 %r21, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 21), align 2497  store i16 %r22, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 22), align 2498  store i16 %r23, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 23), align 2499  store i16 %r24, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 2500  store i16 %r25, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 25), align 2501  store i16 %r26, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 26), align 2502  store i16 %r27, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 27), align 2503  store i16 %r28, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 28), align 2504  store i16 %r29, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 29), align 2505  store i16 %r30, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 30), align 2506  store i16 %r31, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 31), align 2507  ret void508}509 510define void @smul_v64i8() {511; SSE-LABEL: @smul_v64i8(512; SSE-NEXT:    [[TMP1:%.*]] = load <16 x i8>, ptr @a8, align 1513; SSE-NEXT:    [[TMP2:%.*]] = load <16 x i8>, ptr @b8, align 1514; SSE-NEXT:    [[TMP3:%.*]] = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> [[TMP1]], <16 x i8> [[TMP2]], i32 3)515; SSE-NEXT:    store <16 x i8> [[TMP3]], ptr @c8, align 1516; SSE-NEXT:    [[TMP4:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1517; SSE-NEXT:    [[TMP5:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1518; SSE-NEXT:    [[TMP6:%.*]] = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> [[TMP4]], <16 x i8> [[TMP5]], i32 3)519; SSE-NEXT:    store <16 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1520; SSE-NEXT:    [[TMP7:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1521; SSE-NEXT:    [[TMP8:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1522; SSE-NEXT:    [[TMP9:%.*]] = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> [[TMP7]], <16 x i8> [[TMP8]], i32 3)523; SSE-NEXT:    store <16 x i8> [[TMP9]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1524; SSE-NEXT:    [[TMP10:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1525; SSE-NEXT:    [[TMP11:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1526; SSE-NEXT:    [[TMP12:%.*]] = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> [[TMP10]], <16 x i8> [[TMP11]], i32 3)527; SSE-NEXT:    store <16 x i8> [[TMP12]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1528; SSE-NEXT:    ret void529;530; SLM-LABEL: @smul_v64i8(531; SLM-NEXT:    [[TMP1:%.*]] = load <16 x i8>, ptr @a8, align 1532; SLM-NEXT:    [[TMP2:%.*]] = load <16 x i8>, ptr @b8, align 1533; SLM-NEXT:    [[TMP3:%.*]] = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> [[TMP1]], <16 x i8> [[TMP2]], i32 3)534; SLM-NEXT:    store <16 x i8> [[TMP3]], ptr @c8, align 1535; SLM-NEXT:    [[TMP4:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1536; SLM-NEXT:    [[TMP5:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1537; SLM-NEXT:    [[TMP6:%.*]] = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> [[TMP4]], <16 x i8> [[TMP5]], i32 3)538; SLM-NEXT:    store <16 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1539; SLM-NEXT:    [[TMP7:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1540; SLM-NEXT:    [[TMP8:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1541; SLM-NEXT:    [[TMP9:%.*]] = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> [[TMP7]], <16 x i8> [[TMP8]], i32 3)542; SLM-NEXT:    store <16 x i8> [[TMP9]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1543; SLM-NEXT:    [[TMP10:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1544; SLM-NEXT:    [[TMP11:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1545; SLM-NEXT:    [[TMP12:%.*]] = call <16 x i8> @llvm.smul.fix.v16i8(<16 x i8> [[TMP10]], <16 x i8> [[TMP11]], i32 3)546; SLM-NEXT:    store <16 x i8> [[TMP12]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1547; SLM-NEXT:    ret void548;549; AVX-LABEL: @smul_v64i8(550; AVX-NEXT:    [[TMP1:%.*]] = load <32 x i8>, ptr @a8, align 1551; AVX-NEXT:    [[TMP2:%.*]] = load <32 x i8>, ptr @b8, align 1552; AVX-NEXT:    [[TMP3:%.*]] = call <32 x i8> @llvm.smul.fix.v32i8(<32 x i8> [[TMP1]], <32 x i8> [[TMP2]], i32 3)553; AVX-NEXT:    store <32 x i8> [[TMP3]], ptr @c8, align 1554; AVX-NEXT:    [[TMP4:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1555; AVX-NEXT:    [[TMP5:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1556; AVX-NEXT:    [[TMP6:%.*]] = call <32 x i8> @llvm.smul.fix.v32i8(<32 x i8> [[TMP4]], <32 x i8> [[TMP5]], i32 3)557; AVX-NEXT:    store <32 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1558; AVX-NEXT:    ret void559;560; AVX512-LABEL: @smul_v64i8(561; AVX512-NEXT:    [[TMP1:%.*]] = load <64 x i8>, ptr @a8, align 1562; AVX512-NEXT:    [[TMP2:%.*]] = load <64 x i8>, ptr @b8, align 1563; AVX512-NEXT:    [[TMP3:%.*]] = call <64 x i8> @llvm.smul.fix.v64i8(<64 x i8> [[TMP1]], <64 x i8> [[TMP2]], i32 3)564; AVX512-NEXT:    store <64 x i8> [[TMP3]], ptr @c8, align 1565; AVX512-NEXT:    ret void566;567  %a0  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 0 ), align 1568  %a1  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 1 ), align 1569  %a2  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 2 ), align 1570  %a3  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 3 ), align 1571  %a4  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 4 ), align 1572  %a5  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 5 ), align 1573  %a6  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 6 ), align 1574  %a7  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 7 ), align 1575  %a8  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 8 ), align 1576  %a9  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 9 ), align 1577  %a10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 10), align 1578  %a11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 11), align 1579  %a12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 12), align 1580  %a13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 13), align 1581  %a14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 14), align 1582  %a15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 15), align 1583  %a16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1584  %a17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 17), align 1585  %a18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 18), align 1586  %a19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 19), align 1587  %a20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 20), align 1588  %a21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 21), align 1589  %a22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 22), align 1590  %a23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 23), align 1591  %a24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 24), align 1592  %a25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 25), align 1593  %a26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 26), align 1594  %a27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 27), align 1595  %a28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 28), align 1596  %a29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 29), align 1597  %a30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 30), align 1598  %a31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 31), align 1599  %a32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1600  %a33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 33), align 1601  %a34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 34), align 1602  %a35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 35), align 1603  %a36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 36), align 1604  %a37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 37), align 1605  %a38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 38), align 1606  %a39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 39), align 1607  %a40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 40), align 1608  %a41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 41), align 1609  %a42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 42), align 1610  %a43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 43), align 1611  %a44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 44), align 1612  %a45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 45), align 1613  %a46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 46), align 1614  %a47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 47), align 1615  %a48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1616  %a49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 49), align 1617  %a50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 50), align 1618  %a51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 51), align 1619  %a52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 52), align 1620  %a53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 53), align 1621  %a54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 54), align 1622  %a55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 55), align 1623  %a56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 56), align 1624  %a57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 57), align 1625  %a58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 58), align 1626  %a59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 59), align 1627  %a60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 60), align 1628  %a61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 61), align 1629  %a62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 62), align 1630  %a63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 63), align 1631  %b0  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 0 ), align 1632  %b1  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 1 ), align 1633  %b2  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 2 ), align 1634  %b3  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 3 ), align 1635  %b4  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 4 ), align 1636  %b5  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 5 ), align 1637  %b6  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 6 ), align 1638  %b7  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 7 ), align 1639  %b8  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 8 ), align 1640  %b9  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 9 ), align 1641  %b10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 10), align 1642  %b11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 11), align 1643  %b12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 12), align 1644  %b13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 13), align 1645  %b14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 14), align 1646  %b15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 15), align 1647  %b16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1648  %b17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 17), align 1649  %b18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 18), align 1650  %b19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 19), align 1651  %b20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 20), align 1652  %b21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 21), align 1653  %b22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 22), align 1654  %b23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 23), align 1655  %b24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 24), align 1656  %b25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 25), align 1657  %b26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 26), align 1658  %b27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 27), align 1659  %b28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 28), align 1660  %b29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 29), align 1661  %b30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 30), align 1662  %b31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 31), align 1663  %b32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1664  %b33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 33), align 1665  %b34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 34), align 1666  %b35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 35), align 1667  %b36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 36), align 1668  %b37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 37), align 1669  %b38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 38), align 1670  %b39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 39), align 1671  %b40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 40), align 1672  %b41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 41), align 1673  %b42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 42), align 1674  %b43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 43), align 1675  %b44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 44), align 1676  %b45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 45), align 1677  %b46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 46), align 1678  %b47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 47), align 1679  %b48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1680  %b49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 49), align 1681  %b50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 50), align 1682  %b51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 51), align 1683  %b52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 52), align 1684  %b53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 53), align 1685  %b54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 54), align 1686  %b55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 55), align 1687  %b56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 56), align 1688  %b57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 57), align 1689  %b58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 58), align 1690  %b59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 59), align 1691  %b60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 60), align 1692  %b61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 61), align 1693  %b62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 62), align 1694  %b63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 63), align 1695  %r0  = call i8 @llvm.smul.fix.i8(i8 %a0 , i8 %b0 , i32 3)696  %r1  = call i8 @llvm.smul.fix.i8(i8 %a1 , i8 %b1 , i32 3)697  %r2  = call i8 @llvm.smul.fix.i8(i8 %a2 , i8 %b2 , i32 3)698  %r3  = call i8 @llvm.smul.fix.i8(i8 %a3 , i8 %b3 , i32 3)699  %r4  = call i8 @llvm.smul.fix.i8(i8 %a4 , i8 %b4 , i32 3)700  %r5  = call i8 @llvm.smul.fix.i8(i8 %a5 , i8 %b5 , i32 3)701  %r6  = call i8 @llvm.smul.fix.i8(i8 %a6 , i8 %b6 , i32 3)702  %r7  = call i8 @llvm.smul.fix.i8(i8 %a7 , i8 %b7 , i32 3)703  %r8  = call i8 @llvm.smul.fix.i8(i8 %a8 , i8 %b8 , i32 3)704  %r9  = call i8 @llvm.smul.fix.i8(i8 %a9 , i8 %b9 , i32 3)705  %r10 = call i8 @llvm.smul.fix.i8(i8 %a10, i8 %b10, i32 3)706  %r11 = call i8 @llvm.smul.fix.i8(i8 %a11, i8 %b11, i32 3)707  %r12 = call i8 @llvm.smul.fix.i8(i8 %a12, i8 %b12, i32 3)708  %r13 = call i8 @llvm.smul.fix.i8(i8 %a13, i8 %b13, i32 3)709  %r14 = call i8 @llvm.smul.fix.i8(i8 %a14, i8 %b14, i32 3)710  %r15 = call i8 @llvm.smul.fix.i8(i8 %a15, i8 %b15, i32 3)711  %r16 = call i8 @llvm.smul.fix.i8(i8 %a16, i8 %b16, i32 3)712  %r17 = call i8 @llvm.smul.fix.i8(i8 %a17, i8 %b17, i32 3)713  %r18 = call i8 @llvm.smul.fix.i8(i8 %a18, i8 %b18, i32 3)714  %r19 = call i8 @llvm.smul.fix.i8(i8 %a19, i8 %b19, i32 3)715  %r20 = call i8 @llvm.smul.fix.i8(i8 %a20, i8 %b20, i32 3)716  %r21 = call i8 @llvm.smul.fix.i8(i8 %a21, i8 %b21, i32 3)717  %r22 = call i8 @llvm.smul.fix.i8(i8 %a22, i8 %b22, i32 3)718  %r23 = call i8 @llvm.smul.fix.i8(i8 %a23, i8 %b23, i32 3)719  %r24 = call i8 @llvm.smul.fix.i8(i8 %a24, i8 %b24, i32 3)720  %r25 = call i8 @llvm.smul.fix.i8(i8 %a25, i8 %b25, i32 3)721  %r26 = call i8 @llvm.smul.fix.i8(i8 %a26, i8 %b26, i32 3)722  %r27 = call i8 @llvm.smul.fix.i8(i8 %a27, i8 %b27, i32 3)723  %r28 = call i8 @llvm.smul.fix.i8(i8 %a28, i8 %b28, i32 3)724  %r29 = call i8 @llvm.smul.fix.i8(i8 %a29, i8 %b29, i32 3)725  %r30 = call i8 @llvm.smul.fix.i8(i8 %a30, i8 %b30, i32 3)726  %r31 = call i8 @llvm.smul.fix.i8(i8 %a31, i8 %b31, i32 3)727  %r32 = call i8 @llvm.smul.fix.i8(i8 %a32, i8 %b32, i32 3)728  %r33 = call i8 @llvm.smul.fix.i8(i8 %a33, i8 %b33, i32 3)729  %r34 = call i8 @llvm.smul.fix.i8(i8 %a34, i8 %b34, i32 3)730  %r35 = call i8 @llvm.smul.fix.i8(i8 %a35, i8 %b35, i32 3)731  %r36 = call i8 @llvm.smul.fix.i8(i8 %a36, i8 %b36, i32 3)732  %r37 = call i8 @llvm.smul.fix.i8(i8 %a37, i8 %b37, i32 3)733  %r38 = call i8 @llvm.smul.fix.i8(i8 %a38, i8 %b38, i32 3)734  %r39 = call i8 @llvm.smul.fix.i8(i8 %a39, i8 %b39, i32 3)735  %r40 = call i8 @llvm.smul.fix.i8(i8 %a40, i8 %b40, i32 3)736  %r41 = call i8 @llvm.smul.fix.i8(i8 %a41, i8 %b41, i32 3)737  %r42 = call i8 @llvm.smul.fix.i8(i8 %a42, i8 %b42, i32 3)738  %r43 = call i8 @llvm.smul.fix.i8(i8 %a43, i8 %b43, i32 3)739  %r44 = call i8 @llvm.smul.fix.i8(i8 %a44, i8 %b44, i32 3)740  %r45 = call i8 @llvm.smul.fix.i8(i8 %a45, i8 %b45, i32 3)741  %r46 = call i8 @llvm.smul.fix.i8(i8 %a46, i8 %b46, i32 3)742  %r47 = call i8 @llvm.smul.fix.i8(i8 %a47, i8 %b47, i32 3)743  %r48 = call i8 @llvm.smul.fix.i8(i8 %a48, i8 %b48, i32 3)744  %r49 = call i8 @llvm.smul.fix.i8(i8 %a49, i8 %b49, i32 3)745  %r50 = call i8 @llvm.smul.fix.i8(i8 %a50, i8 %b50, i32 3)746  %r51 = call i8 @llvm.smul.fix.i8(i8 %a51, i8 %b51, i32 3)747  %r52 = call i8 @llvm.smul.fix.i8(i8 %a52, i8 %b52, i32 3)748  %r53 = call i8 @llvm.smul.fix.i8(i8 %a53, i8 %b53, i32 3)749  %r54 = call i8 @llvm.smul.fix.i8(i8 %a54, i8 %b54, i32 3)750  %r55 = call i8 @llvm.smul.fix.i8(i8 %a55, i8 %b55, i32 3)751  %r56 = call i8 @llvm.smul.fix.i8(i8 %a56, i8 %b56, i32 3)752  %r57 = call i8 @llvm.smul.fix.i8(i8 %a57, i8 %b57, i32 3)753  %r58 = call i8 @llvm.smul.fix.i8(i8 %a58, i8 %b58, i32 3)754  %r59 = call i8 @llvm.smul.fix.i8(i8 %a59, i8 %b59, i32 3)755  %r60 = call i8 @llvm.smul.fix.i8(i8 %a60, i8 %b60, i32 3)756  %r61 = call i8 @llvm.smul.fix.i8(i8 %a61, i8 %b61, i32 3)757  %r62 = call i8 @llvm.smul.fix.i8(i8 %a62, i8 %b62, i32 3)758  %r63 = call i8 @llvm.smul.fix.i8(i8 %a63, i8 %b63, i32 3)759  store i8 %r0 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 0 ), align 1760  store i8 %r1 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 1 ), align 1761  store i8 %r2 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 2 ), align 1762  store i8 %r3 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 3 ), align 1763  store i8 %r4 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 4 ), align 1764  store i8 %r5 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 5 ), align 1765  store i8 %r6 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 6 ), align 1766  store i8 %r7 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 7 ), align 1767  store i8 %r8 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 8 ), align 1768  store i8 %r9 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 9 ), align 1769  store i8 %r10, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 10), align 1770  store i8 %r11, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 11), align 1771  store i8 %r12, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 12), align 1772  store i8 %r13, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 13), align 1773  store i8 %r14, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 14), align 1774  store i8 %r15, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 15), align 1775  store i8 %r16, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 1776  store i8 %r17, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 17), align 1777  store i8 %r18, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 18), align 1778  store i8 %r19, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 19), align 1779  store i8 %r20, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 20), align 1780  store i8 %r21, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 21), align 1781  store i8 %r22, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 22), align 1782  store i8 %r23, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 23), align 1783  store i8 %r24, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 24), align 1784  store i8 %r25, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 25), align 1785  store i8 %r26, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 26), align 1786  store i8 %r27, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 27), align 1787  store i8 %r28, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 28), align 1788  store i8 %r29, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 29), align 1789  store i8 %r30, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 30), align 1790  store i8 %r31, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 31), align 1791  store i8 %r32, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 1792  store i8 %r33, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 33), align 1793  store i8 %r34, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 34), align 1794  store i8 %r35, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 35), align 1795  store i8 %r36, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 36), align 1796  store i8 %r37, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 37), align 1797  store i8 %r38, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 38), align 1798  store i8 %r39, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 39), align 1799  store i8 %r40, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 40), align 1800  store i8 %r41, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 41), align 1801  store i8 %r42, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 42), align 1802  store i8 %r43, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 43), align 1803  store i8 %r44, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 44), align 1804  store i8 %r45, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 45), align 1805  store i8 %r46, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 46), align 1806  store i8 %r47, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 47), align 1807  store i8 %r48, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 1808  store i8 %r49, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 49), align 1809  store i8 %r50, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 50), align 1810  store i8 %r51, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 51), align 1811  store i8 %r52, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 52), align 1812  store i8 %r53, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 53), align 1813  store i8 %r54, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 54), align 1814  store i8 %r55, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 55), align 1815  store i8 %r56, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 56), align 1816  store i8 %r57, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 57), align 1817  store i8 %r58, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 58), align 1818  store i8 %r59, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 59), align 1819  store i8 %r60, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 60), align 1820  store i8 %r61, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 61), align 1821  store i8 %r62, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 62), align 1822  store i8 %r63, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 63), align 1823  ret void824}825 826declare i64 @llvm.umul.fix.i64(i64, i64, i32)827declare i32 @llvm.umul.fix.i32(i32, i32, i32)828declare i16 @llvm.umul.fix.i16(i16, i16, i32)829declare i8  @llvm.umul.fix.i8 (i8 , i8 , i32)830 831define void @umul_v8i64() {832; SSE-LABEL: @umul_v8i64(833; SSE-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @a64, align 8834; SSE-NEXT:    [[TMP2:%.*]] = load <2 x i64>, ptr @b64, align 8835; SSE-NEXT:    [[TMP3:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP1]], <2 x i64> [[TMP2]], i32 3)836; SSE-NEXT:    store <2 x i64> [[TMP3]], ptr @c64, align 8837; SSE-NEXT:    [[TMP4:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8838; SSE-NEXT:    [[TMP5:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8839; SSE-NEXT:    [[TMP6:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP4]], <2 x i64> [[TMP5]], i32 3)840; SSE-NEXT:    store <2 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8841; SSE-NEXT:    [[TMP7:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8842; SSE-NEXT:    [[TMP8:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8843; SSE-NEXT:    [[TMP9:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP7]], <2 x i64> [[TMP8]], i32 3)844; SSE-NEXT:    store <2 x i64> [[TMP9]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8845; SSE-NEXT:    [[TMP10:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8846; SSE-NEXT:    [[TMP11:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8847; SSE-NEXT:    [[TMP12:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP10]], <2 x i64> [[TMP11]], i32 3)848; SSE-NEXT:    store <2 x i64> [[TMP12]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8849; SSE-NEXT:    ret void850;851; SLM-LABEL: @umul_v8i64(852; SLM-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @a64, align 8853; SLM-NEXT:    [[TMP2:%.*]] = load <2 x i64>, ptr @b64, align 8854; SLM-NEXT:    [[TMP3:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP1]], <2 x i64> [[TMP2]], i32 3)855; SLM-NEXT:    store <2 x i64> [[TMP3]], ptr @c64, align 8856; SLM-NEXT:    [[TMP4:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8857; SLM-NEXT:    [[TMP5:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8858; SLM-NEXT:    [[TMP6:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP4]], <2 x i64> [[TMP5]], i32 3)859; SLM-NEXT:    store <2 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8860; SLM-NEXT:    [[TMP7:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8861; SLM-NEXT:    [[TMP8:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8862; SLM-NEXT:    [[TMP9:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP7]], <2 x i64> [[TMP8]], i32 3)863; SLM-NEXT:    store <2 x i64> [[TMP9]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8864; SLM-NEXT:    [[TMP10:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8865; SLM-NEXT:    [[TMP11:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8866; SLM-NEXT:    [[TMP12:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP10]], <2 x i64> [[TMP11]], i32 3)867; SLM-NEXT:    store <2 x i64> [[TMP12]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8868; SLM-NEXT:    ret void869;870; AVX1-LABEL: @umul_v8i64(871; AVX1-NEXT:    [[TMP1:%.*]] = load <2 x i64>, ptr @a64, align 8872; AVX1-NEXT:    [[TMP2:%.*]] = load <2 x i64>, ptr @b64, align 8873; AVX1-NEXT:    [[TMP3:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP1]], <2 x i64> [[TMP2]], i32 3)874; AVX1-NEXT:    store <2 x i64> [[TMP3]], ptr @c64, align 8875; AVX1-NEXT:    [[TMP4:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8876; AVX1-NEXT:    [[TMP5:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8877; AVX1-NEXT:    [[TMP6:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP4]], <2 x i64> [[TMP5]], i32 3)878; AVX1-NEXT:    store <2 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8879; AVX1-NEXT:    [[TMP7:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8880; AVX1-NEXT:    [[TMP8:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8881; AVX1-NEXT:    [[TMP9:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP7]], <2 x i64> [[TMP8]], i32 3)882; AVX1-NEXT:    store <2 x i64> [[TMP9]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8883; AVX1-NEXT:    [[TMP10:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8884; AVX1-NEXT:    [[TMP11:%.*]] = load <2 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8885; AVX1-NEXT:    [[TMP12:%.*]] = call <2 x i64> @llvm.umul.fix.v2i64(<2 x i64> [[TMP10]], <2 x i64> [[TMP11]], i32 3)886; AVX1-NEXT:    store <2 x i64> [[TMP12]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8887; AVX1-NEXT:    ret void888;889; AVX2-LABEL: @umul_v8i64(890; AVX2-NEXT:    [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 8891; AVX2-NEXT:    [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 8892; AVX2-NEXT:    [[TMP3:%.*]] = call <4 x i64> @llvm.umul.fix.v4i64(<4 x i64> [[TMP1]], <4 x i64> [[TMP2]], i32 3)893; AVX2-NEXT:    store <4 x i64> [[TMP3]], ptr @c64, align 8894; AVX2-NEXT:    [[TMP4:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8895; AVX2-NEXT:    [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8896; AVX2-NEXT:    [[TMP6:%.*]] = call <4 x i64> @llvm.umul.fix.v4i64(<4 x i64> [[TMP4]], <4 x i64> [[TMP5]], i32 3)897; AVX2-NEXT:    store <4 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8898; AVX2-NEXT:    ret void899;900; AVX512-LABEL: @umul_v8i64(901; AVX512-NEXT:    [[TMP1:%.*]] = load <8 x i64>, ptr @a64, align 8902; AVX512-NEXT:    [[TMP2:%.*]] = load <8 x i64>, ptr @b64, align 8903; AVX512-NEXT:    [[TMP3:%.*]] = call <8 x i64> @llvm.umul.fix.v8i64(<8 x i64> [[TMP1]], <8 x i64> [[TMP2]], i32 3)904; AVX512-NEXT:    store <8 x i64> [[TMP3]], ptr @c64, align 8905; AVX512-NEXT:    ret void906;907; AVX256BW-LABEL: @umul_v8i64(908; AVX256BW-NEXT:    [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 8909; AVX256BW-NEXT:    [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 8910; AVX256BW-NEXT:    [[TMP3:%.*]] = call <4 x i64> @llvm.umul.fix.v4i64(<4 x i64> [[TMP1]], <4 x i64> [[TMP2]], i32 3)911; AVX256BW-NEXT:    store <4 x i64> [[TMP3]], ptr @c64, align 8912; AVX256BW-NEXT:    [[TMP4:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8913; AVX256BW-NEXT:    [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8914; AVX256BW-NEXT:    [[TMP6:%.*]] = call <4 x i64> @llvm.umul.fix.v4i64(<4 x i64> [[TMP4]], <4 x i64> [[TMP5]], i32 3)915; AVX256BW-NEXT:    store <4 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8916; AVX256BW-NEXT:    ret void917;918  %a0 = load i64, ptr @a64, align 8919  %a1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 8920  %a2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8921  %a3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 8922  %a4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8923  %a5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 8924  %a6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8925  %a7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 8926  %b0 = load i64, ptr @b64, align 8927  %b1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 8928  %b2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8929  %b3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 8930  %b4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8931  %b5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 8932  %b6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8933  %b7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 8934  %r0 = call i64 @llvm.umul.fix.i64(i64 %a0, i64 %b0, i32 3)935  %r1 = call i64 @llvm.umul.fix.i64(i64 %a1, i64 %b1, i32 3)936  %r2 = call i64 @llvm.umul.fix.i64(i64 %a2, i64 %b2, i32 3)937  %r3 = call i64 @llvm.umul.fix.i64(i64 %a3, i64 %b3, i32 3)938  %r4 = call i64 @llvm.umul.fix.i64(i64 %a4, i64 %b4, i32 3)939  %r5 = call i64 @llvm.umul.fix.i64(i64 %a5, i64 %b5, i32 3)940  %r6 = call i64 @llvm.umul.fix.i64(i64 %a6, i64 %b6, i32 3)941  %r7 = call i64 @llvm.umul.fix.i64(i64 %a7, i64 %b7, i32 3)942  store i64 %r0, ptr @c64, align 8943  store i64 %r1, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 1), align 8944  store i64 %r2, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 2), align 8945  store i64 %r3, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 3), align 8946  store i64 %r4, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 4), align 8947  store i64 %r5, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 5), align 8948  store i64 %r6, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 6), align 8949  store i64 %r7, ptr getelementptr inbounds ([8 x i64], ptr @c64, i32 0, i64 7), align 8950  ret void951}952 953define void @umul_v16i32() {954; SSE-LABEL: @umul_v16i32(955; SSE-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr @a32, align 4956; SSE-NEXT:    [[TMP2:%.*]] = load <4 x i32>, ptr @b32, align 4957; SSE-NEXT:    [[TMP3:%.*]] = call <4 x i32> @llvm.umul.fix.v4i32(<4 x i32> [[TMP1]], <4 x i32> [[TMP2]], i32 3)958; SSE-NEXT:    store <4 x i32> [[TMP3]], ptr @c32, align 4959; SSE-NEXT:    [[TMP4:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4960; SSE-NEXT:    [[TMP5:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4961; SSE-NEXT:    [[TMP6:%.*]] = call <4 x i32> @llvm.umul.fix.v4i32(<4 x i32> [[TMP4]], <4 x i32> [[TMP5]], i32 3)962; SSE-NEXT:    store <4 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4), align 4963; SSE-NEXT:    [[TMP7:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4964; SSE-NEXT:    [[TMP8:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4965; SSE-NEXT:    [[TMP9:%.*]] = call <4 x i32> @llvm.umul.fix.v4i32(<4 x i32> [[TMP7]], <4 x i32> [[TMP8]], i32 3)966; SSE-NEXT:    store <4 x i32> [[TMP9]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 4967; SSE-NEXT:    [[TMP10:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4968; SSE-NEXT:    [[TMP11:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4969; SSE-NEXT:    [[TMP12:%.*]] = call <4 x i32> @llvm.umul.fix.v4i32(<4 x i32> [[TMP10]], <4 x i32> [[TMP11]], i32 3)970; SSE-NEXT:    store <4 x i32> [[TMP12]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 4971; SSE-NEXT:    ret void972;973; SLM-LABEL: @umul_v16i32(974; SLM-NEXT:    [[A0:%.*]] = load i32, ptr @a32, align 4975; SLM-NEXT:    [[A1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1), align 4976; SLM-NEXT:    [[A2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2), align 4977; SLM-NEXT:    [[A3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3), align 4978; SLM-NEXT:    [[A4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4979; SLM-NEXT:    [[A5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5), align 4980; SLM-NEXT:    [[A6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6), align 4981; SLM-NEXT:    [[A7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7), align 4982; SLM-NEXT:    [[A8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4983; SLM-NEXT:    [[A9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9), align 4984; SLM-NEXT:    [[A10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 4985; SLM-NEXT:    [[A11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 4986; SLM-NEXT:    [[A12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4987; SLM-NEXT:    [[A13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 4988; SLM-NEXT:    [[A14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 4989; SLM-NEXT:    [[A15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 4990; SLM-NEXT:    [[B0:%.*]] = load i32, ptr @b32, align 4991; SLM-NEXT:    [[B1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1), align 4992; SLM-NEXT:    [[B2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2), align 4993; SLM-NEXT:    [[B3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3), align 4994; SLM-NEXT:    [[B4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4995; SLM-NEXT:    [[B5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5), align 4996; SLM-NEXT:    [[B6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6), align 4997; SLM-NEXT:    [[B7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7), align 4998; SLM-NEXT:    [[B8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4999; SLM-NEXT:    [[B9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9), align 41000; SLM-NEXT:    [[B10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 41001; SLM-NEXT:    [[B11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 41002; SLM-NEXT:    [[B12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 41003; SLM-NEXT:    [[B13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 41004; SLM-NEXT:    [[B14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 41005; SLM-NEXT:    [[B15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 41006; SLM-NEXT:    [[R0:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A0]], i32 [[B0]], i32 3)1007; SLM-NEXT:    [[R1:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A1]], i32 [[B1]], i32 3)1008; SLM-NEXT:    [[R2:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A2]], i32 [[B2]], i32 3)1009; SLM-NEXT:    [[R3:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A3]], i32 [[B3]], i32 3)1010; SLM-NEXT:    [[R4:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A4]], i32 [[B4]], i32 3)1011; SLM-NEXT:    [[R5:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A5]], i32 [[B5]], i32 3)1012; SLM-NEXT:    [[R6:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A6]], i32 [[B6]], i32 3)1013; SLM-NEXT:    [[R7:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A7]], i32 [[B7]], i32 3)1014; SLM-NEXT:    [[R8:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A8]], i32 [[B8]], i32 3)1015; SLM-NEXT:    [[R9:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A9]], i32 [[B9]], i32 3)1016; SLM-NEXT:    [[R10:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A10]], i32 [[B10]], i32 3)1017; SLM-NEXT:    [[R11:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A11]], i32 [[B11]], i32 3)1018; SLM-NEXT:    [[R12:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A12]], i32 [[B12]], i32 3)1019; SLM-NEXT:    [[R13:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A13]], i32 [[B13]], i32 3)1020; SLM-NEXT:    [[R14:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A14]], i32 [[B14]], i32 3)1021; SLM-NEXT:    [[R15:%.*]] = call i32 @llvm.umul.fix.i32(i32 [[A15]], i32 [[B15]], i32 3)1022; SLM-NEXT:    store i32 [[R0]], ptr @c32, align 41023; SLM-NEXT:    store i32 [[R1]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 1), align 41024; SLM-NEXT:    store i32 [[R2]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 2), align 41025; SLM-NEXT:    store i32 [[R3]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 3), align 41026; SLM-NEXT:    store i32 [[R4]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4), align 41027; SLM-NEXT:    store i32 [[R5]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 5), align 41028; SLM-NEXT:    store i32 [[R6]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 6), align 41029; SLM-NEXT:    store i32 [[R7]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 7), align 41030; SLM-NEXT:    store i32 [[R8]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 41031; SLM-NEXT:    store i32 [[R9]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 9), align 41032; SLM-NEXT:    store i32 [[R10]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 10), align 41033; SLM-NEXT:    store i32 [[R11]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 11), align 41034; SLM-NEXT:    store i32 [[R12]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 41035; SLM-NEXT:    store i32 [[R13]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 13), align 41036; SLM-NEXT:    store i32 [[R14]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 14), align 41037; SLM-NEXT:    store i32 [[R15]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 15), align 41038; SLM-NEXT:    ret void1039;1040; AVX-LABEL: @umul_v16i32(1041; AVX-NEXT:    [[TMP1:%.*]] = load <8 x i32>, ptr @a32, align 41042; AVX-NEXT:    [[TMP2:%.*]] = load <8 x i32>, ptr @b32, align 41043; AVX-NEXT:    [[TMP3:%.*]] = call <8 x i32> @llvm.umul.fix.v8i32(<8 x i32> [[TMP1]], <8 x i32> [[TMP2]], i32 3)1044; AVX-NEXT:    store <8 x i32> [[TMP3]], ptr @c32, align 41045; AVX-NEXT:    [[TMP4:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 41046; AVX-NEXT:    [[TMP5:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 41047; AVX-NEXT:    [[TMP6:%.*]] = call <8 x i32> @llvm.umul.fix.v8i32(<8 x i32> [[TMP4]], <8 x i32> [[TMP5]], i32 3)1048; AVX-NEXT:    store <8 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8), align 41049; AVX-NEXT:    ret void1050;1051; AVX512-LABEL: @umul_v16i32(1052; AVX512-NEXT:    [[TMP1:%.*]] = load <16 x i32>, ptr @a32, align 41053; AVX512-NEXT:    [[TMP2:%.*]] = load <16 x i32>, ptr @b32, align 41054; AVX512-NEXT:    [[TMP3:%.*]] = call <16 x i32> @llvm.umul.fix.v16i32(<16 x i32> [[TMP1]], <16 x i32> [[TMP2]], i32 3)1055; AVX512-NEXT:    store <16 x i32> [[TMP3]], ptr @c32, align 41056; AVX512-NEXT:    ret void1057;1058  %a0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 0 ), align 41059  %a1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1 ), align 41060  %a2  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2 ), align 41061  %a3  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3 ), align 41062  %a4  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4 ), align 41063  %a5  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5 ), align 41064  %a6  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6 ), align 41065  %a7  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7 ), align 41066  %a8  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8 ), align 41067  %a9  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9 ), align 41068  %a10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 41069  %a11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 41070  %a12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 41071  %a13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 41072  %a14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 41073  %a15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 41074  %b0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 0 ), align 41075  %b1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1 ), align 41076  %b2  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2 ), align 41077  %b3  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3 ), align 41078  %b4  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4 ), align 41079  %b5  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5 ), align 41080  %b6  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6 ), align 41081  %b7  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7 ), align 41082  %b8  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8 ), align 41083  %b9  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9 ), align 41084  %b10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 41085  %b11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 41086  %b12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 41087  %b13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 41088  %b14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 41089  %b15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 41090  %r0  = call i32 @llvm.umul.fix.i32(i32 %a0 , i32 %b0 , i32 3)1091  %r1  = call i32 @llvm.umul.fix.i32(i32 %a1 , i32 %b1 , i32 3)1092  %r2  = call i32 @llvm.umul.fix.i32(i32 %a2 , i32 %b2 , i32 3)1093  %r3  = call i32 @llvm.umul.fix.i32(i32 %a3 , i32 %b3 , i32 3)1094  %r4  = call i32 @llvm.umul.fix.i32(i32 %a4 , i32 %b4 , i32 3)1095  %r5  = call i32 @llvm.umul.fix.i32(i32 %a5 , i32 %b5 , i32 3)1096  %r6  = call i32 @llvm.umul.fix.i32(i32 %a6 , i32 %b6 , i32 3)1097  %r7  = call i32 @llvm.umul.fix.i32(i32 %a7 , i32 %b7 , i32 3)1098  %r8  = call i32 @llvm.umul.fix.i32(i32 %a8 , i32 %b8 , i32 3)1099  %r9  = call i32 @llvm.umul.fix.i32(i32 %a9 , i32 %b9 , i32 3)1100  %r10 = call i32 @llvm.umul.fix.i32(i32 %a10, i32 %b10, i32 3)1101  %r11 = call i32 @llvm.umul.fix.i32(i32 %a11, i32 %b11, i32 3)1102  %r12 = call i32 @llvm.umul.fix.i32(i32 %a12, i32 %b12, i32 3)1103  %r13 = call i32 @llvm.umul.fix.i32(i32 %a13, i32 %b13, i32 3)1104  %r14 = call i32 @llvm.umul.fix.i32(i32 %a14, i32 %b14, i32 3)1105  %r15 = call i32 @llvm.umul.fix.i32(i32 %a15, i32 %b15, i32 3)1106  store i32 %r0 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 0 ), align 41107  store i32 %r1 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 1 ), align 41108  store i32 %r2 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 2 ), align 41109  store i32 %r3 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 3 ), align 41110  store i32 %r4 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 4 ), align 41111  store i32 %r5 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 5 ), align 41112  store i32 %r6 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 6 ), align 41113  store i32 %r7 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 7 ), align 41114  store i32 %r8 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 8 ), align 41115  store i32 %r9 , ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 9 ), align 41116  store i32 %r10, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 10), align 41117  store i32 %r11, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 11), align 41118  store i32 %r12, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 12), align 41119  store i32 %r13, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 13), align 41120  store i32 %r14, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 14), align 41121  store i32 %r15, ptr getelementptr inbounds ([16 x i32], ptr @c32, i32 0, i64 15), align 41122  ret void1123}1124 1125define void @umul_v32i16() {1126; SSE-LABEL: @umul_v32i16(1127; SSE-NEXT:    [[TMP1:%.*]] = load <8 x i16>, ptr @a16, align 21128; SSE-NEXT:    [[TMP2:%.*]] = load <8 x i16>, ptr @b16, align 21129; SSE-NEXT:    [[TMP3:%.*]] = call <8 x i16> @llvm.umul.fix.v8i16(<8 x i16> [[TMP1]], <8 x i16> [[TMP2]], i32 3)1130; SSE-NEXT:    store <8 x i16> [[TMP3]], ptr @c16, align 21131; SSE-NEXT:    [[TMP4:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8), align 21132; SSE-NEXT:    [[TMP5:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8), align 21133; SSE-NEXT:    [[TMP6:%.*]] = call <8 x i16> @llvm.umul.fix.v8i16(<8 x i16> [[TMP4]], <8 x i16> [[TMP5]], i32 3)1134; SSE-NEXT:    store <8 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8), align 21135; SSE-NEXT:    [[TMP7:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 21136; SSE-NEXT:    [[TMP8:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 21137; SSE-NEXT:    [[TMP9:%.*]] = call <8 x i16> @llvm.umul.fix.v8i16(<8 x i16> [[TMP7]], <8 x i16> [[TMP8]], i32 3)1138; SSE-NEXT:    store <8 x i16> [[TMP9]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 21139; SSE-NEXT:    [[TMP10:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 21140; SSE-NEXT:    [[TMP11:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 21141; SSE-NEXT:    [[TMP12:%.*]] = call <8 x i16> @llvm.umul.fix.v8i16(<8 x i16> [[TMP10]], <8 x i16> [[TMP11]], i32 3)1142; SSE-NEXT:    store <8 x i16> [[TMP12]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 21143; SSE-NEXT:    ret void1144;1145; SLM-LABEL: @umul_v32i16(1146; SLM-NEXT:    [[TMP1:%.*]] = load <8 x i16>, ptr @a16, align 21147; SLM-NEXT:    [[TMP2:%.*]] = load <8 x i16>, ptr @b16, align 21148; SLM-NEXT:    [[TMP3:%.*]] = call <8 x i16> @llvm.umul.fix.v8i16(<8 x i16> [[TMP1]], <8 x i16> [[TMP2]], i32 3)1149; SLM-NEXT:    store <8 x i16> [[TMP3]], ptr @c16, align 21150; SLM-NEXT:    [[TMP4:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8), align 21151; SLM-NEXT:    [[TMP5:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8), align 21152; SLM-NEXT:    [[TMP6:%.*]] = call <8 x i16> @llvm.umul.fix.v8i16(<8 x i16> [[TMP4]], <8 x i16> [[TMP5]], i32 3)1153; SLM-NEXT:    store <8 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8), align 21154; SLM-NEXT:    [[TMP7:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 21155; SLM-NEXT:    [[TMP8:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 21156; SLM-NEXT:    [[TMP9:%.*]] = call <8 x i16> @llvm.umul.fix.v8i16(<8 x i16> [[TMP7]], <8 x i16> [[TMP8]], i32 3)1157; SLM-NEXT:    store <8 x i16> [[TMP9]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 21158; SLM-NEXT:    [[TMP10:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 21159; SLM-NEXT:    [[TMP11:%.*]] = load <8 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 21160; SLM-NEXT:    [[TMP12:%.*]] = call <8 x i16> @llvm.umul.fix.v8i16(<8 x i16> [[TMP10]], <8 x i16> [[TMP11]], i32 3)1161; SLM-NEXT:    store <8 x i16> [[TMP12]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 21162; SLM-NEXT:    ret void1163;1164; AVX-LABEL: @umul_v32i16(1165; AVX-NEXT:    [[TMP1:%.*]] = load <16 x i16>, ptr @a16, align 21166; AVX-NEXT:    [[TMP2:%.*]] = load <16 x i16>, ptr @b16, align 21167; AVX-NEXT:    [[TMP3:%.*]] = call <16 x i16> @llvm.umul.fix.v16i16(<16 x i16> [[TMP1]], <16 x i16> [[TMP2]], i32 3)1168; AVX-NEXT:    store <16 x i16> [[TMP3]], ptr @c16, align 21169; AVX-NEXT:    [[TMP4:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 21170; AVX-NEXT:    [[TMP5:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 21171; AVX-NEXT:    [[TMP6:%.*]] = call <16 x i16> @llvm.umul.fix.v16i16(<16 x i16> [[TMP4]], <16 x i16> [[TMP5]], i32 3)1172; AVX-NEXT:    store <16 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 21173; AVX-NEXT:    ret void1174;1175; AVX512-LABEL: @umul_v32i16(1176; AVX512-NEXT:    [[TMP1:%.*]] = load <32 x i16>, ptr @a16, align 21177; AVX512-NEXT:    [[TMP2:%.*]] = load <32 x i16>, ptr @b16, align 21178; AVX512-NEXT:    [[TMP3:%.*]] = call <32 x i16> @llvm.umul.fix.v32i16(<32 x i16> [[TMP1]], <32 x i16> [[TMP2]], i32 3)1179; AVX512-NEXT:    store <32 x i16> [[TMP3]], ptr @c16, align 21180; AVX512-NEXT:    ret void1181;1182  %a0  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 0 ), align 21183  %a1  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 1 ), align 21184  %a2  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 2 ), align 21185  %a3  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 3 ), align 21186  %a4  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 4 ), align 21187  %a5  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 5 ), align 21188  %a6  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 6 ), align 21189  %a7  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 7 ), align 21190  %a8  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8 ), align 21191  %a9  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 9 ), align 21192  %a10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 10), align 21193  %a11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 11), align 21194  %a12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 12), align 21195  %a13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 13), align 21196  %a14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 14), align 21197  %a15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 15), align 21198  %a16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 21199  %a17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 17), align 21200  %a18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 18), align 21201  %a19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 19), align 21202  %a20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 20), align 21203  %a21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 21), align 21204  %a22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 22), align 21205  %a23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 23), align 21206  %a24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 21207  %a25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 25), align 21208  %a26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 26), align 21209  %a27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 27), align 21210  %a28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 28), align 21211  %a29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 29), align 21212  %a30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 30), align 21213  %a31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 31), align 21214  %b0  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 0 ), align 21215  %b1  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 1 ), align 21216  %b2  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 2 ), align 21217  %b3  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 3 ), align 21218  %b4  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 4 ), align 21219  %b5  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 5 ), align 21220  %b6  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 6 ), align 21221  %b7  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 7 ), align 21222  %b8  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8 ), align 21223  %b9  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 9 ), align 21224  %b10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 10), align 21225  %b11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 11), align 21226  %b12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 12), align 21227  %b13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 13), align 21228  %b14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 14), align 21229  %b15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 15), align 21230  %b16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 21231  %b17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 17), align 21232  %b18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 18), align 21233  %b19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 19), align 21234  %b20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 20), align 21235  %b21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 21), align 21236  %b22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 22), align 21237  %b23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 23), align 21238  %b24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 21239  %b25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 25), align 21240  %b26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 26), align 21241  %b27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 27), align 21242  %b28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 28), align 21243  %b29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 29), align 21244  %b30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 30), align 21245  %b31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 31), align 21246  %r0  = call i16 @llvm.umul.fix.i16(i16 %a0 , i16 %b0 , i32 3)1247  %r1  = call i16 @llvm.umul.fix.i16(i16 %a1 , i16 %b1 , i32 3)1248  %r2  = call i16 @llvm.umul.fix.i16(i16 %a2 , i16 %b2 , i32 3)1249  %r3  = call i16 @llvm.umul.fix.i16(i16 %a3 , i16 %b3 , i32 3)1250  %r4  = call i16 @llvm.umul.fix.i16(i16 %a4 , i16 %b4 , i32 3)1251  %r5  = call i16 @llvm.umul.fix.i16(i16 %a5 , i16 %b5 , i32 3)1252  %r6  = call i16 @llvm.umul.fix.i16(i16 %a6 , i16 %b6 , i32 3)1253  %r7  = call i16 @llvm.umul.fix.i16(i16 %a7 , i16 %b7 , i32 3)1254  %r8  = call i16 @llvm.umul.fix.i16(i16 %a8 , i16 %b8 , i32 3)1255  %r9  = call i16 @llvm.umul.fix.i16(i16 %a9 , i16 %b9 , i32 3)1256  %r10 = call i16 @llvm.umul.fix.i16(i16 %a10, i16 %b10, i32 3)1257  %r11 = call i16 @llvm.umul.fix.i16(i16 %a11, i16 %b11, i32 3)1258  %r12 = call i16 @llvm.umul.fix.i16(i16 %a12, i16 %b12, i32 3)1259  %r13 = call i16 @llvm.umul.fix.i16(i16 %a13, i16 %b13, i32 3)1260  %r14 = call i16 @llvm.umul.fix.i16(i16 %a14, i16 %b14, i32 3)1261  %r15 = call i16 @llvm.umul.fix.i16(i16 %a15, i16 %b15, i32 3)1262  %r16 = call i16 @llvm.umul.fix.i16(i16 %a16, i16 %b16, i32 3)1263  %r17 = call i16 @llvm.umul.fix.i16(i16 %a17, i16 %b17, i32 3)1264  %r18 = call i16 @llvm.umul.fix.i16(i16 %a18, i16 %b18, i32 3)1265  %r19 = call i16 @llvm.umul.fix.i16(i16 %a19, i16 %b19, i32 3)1266  %r20 = call i16 @llvm.umul.fix.i16(i16 %a20, i16 %b20, i32 3)1267  %r21 = call i16 @llvm.umul.fix.i16(i16 %a21, i16 %b21, i32 3)1268  %r22 = call i16 @llvm.umul.fix.i16(i16 %a22, i16 %b22, i32 3)1269  %r23 = call i16 @llvm.umul.fix.i16(i16 %a23, i16 %b23, i32 3)1270  %r24 = call i16 @llvm.umul.fix.i16(i16 %a24, i16 %b24, i32 3)1271  %r25 = call i16 @llvm.umul.fix.i16(i16 %a25, i16 %b25, i32 3)1272  %r26 = call i16 @llvm.umul.fix.i16(i16 %a26, i16 %b26, i32 3)1273  %r27 = call i16 @llvm.umul.fix.i16(i16 %a27, i16 %b27, i32 3)1274  %r28 = call i16 @llvm.umul.fix.i16(i16 %a28, i16 %b28, i32 3)1275  %r29 = call i16 @llvm.umul.fix.i16(i16 %a29, i16 %b29, i32 3)1276  %r30 = call i16 @llvm.umul.fix.i16(i16 %a30, i16 %b30, i32 3)1277  %r31 = call i16 @llvm.umul.fix.i16(i16 %a31, i16 %b31, i32 3)1278  store i16 %r0 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 0 ), align 21279  store i16 %r1 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 1 ), align 21280  store i16 %r2 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 2 ), align 21281  store i16 %r3 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 3 ), align 21282  store i16 %r4 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 4 ), align 21283  store i16 %r5 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 5 ), align 21284  store i16 %r6 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 6 ), align 21285  store i16 %r7 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 7 ), align 21286  store i16 %r8 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 8 ), align 21287  store i16 %r9 , ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 9 ), align 21288  store i16 %r10, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 10), align 21289  store i16 %r11, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 11), align 21290  store i16 %r12, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 12), align 21291  store i16 %r13, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 13), align 21292  store i16 %r14, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 14), align 21293  store i16 %r15, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 15), align 21294  store i16 %r16, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 16), align 21295  store i16 %r17, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 17), align 21296  store i16 %r18, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 18), align 21297  store i16 %r19, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 19), align 21298  store i16 %r20, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 20), align 21299  store i16 %r21, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 21), align 21300  store i16 %r22, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 22), align 21301  store i16 %r23, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 23), align 21302  store i16 %r24, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 24), align 21303  store i16 %r25, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 25), align 21304  store i16 %r26, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 26), align 21305  store i16 %r27, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 27), align 21306  store i16 %r28, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 28), align 21307  store i16 %r29, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 29), align 21308  store i16 %r30, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 30), align 21309  store i16 %r31, ptr getelementptr inbounds ([32 x i16], ptr @c16, i32 0, i64 31), align 21310  ret void1311}1312 1313define void @umul_v64i8() {1314; SSE-LABEL: @umul_v64i8(1315; SSE-NEXT:    [[TMP1:%.*]] = load <16 x i8>, ptr @a8, align 11316; SSE-NEXT:    [[TMP2:%.*]] = load <16 x i8>, ptr @b8, align 11317; SSE-NEXT:    [[TMP3:%.*]] = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> [[TMP1]], <16 x i8> [[TMP2]], i32 3)1318; SSE-NEXT:    store <16 x i8> [[TMP3]], ptr @c8, align 11319; SSE-NEXT:    [[TMP4:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 11320; SSE-NEXT:    [[TMP5:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 11321; SSE-NEXT:    [[TMP6:%.*]] = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> [[TMP4]], <16 x i8> [[TMP5]], i32 3)1322; SSE-NEXT:    store <16 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 11323; SSE-NEXT:    [[TMP7:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 11324; SSE-NEXT:    [[TMP8:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 11325; SSE-NEXT:    [[TMP9:%.*]] = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> [[TMP7]], <16 x i8> [[TMP8]], i32 3)1326; SSE-NEXT:    store <16 x i8> [[TMP9]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 11327; SSE-NEXT:    [[TMP10:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 11328; SSE-NEXT:    [[TMP11:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 11329; SSE-NEXT:    [[TMP12:%.*]] = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> [[TMP10]], <16 x i8> [[TMP11]], i32 3)1330; SSE-NEXT:    store <16 x i8> [[TMP12]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 11331; SSE-NEXT:    ret void1332;1333; SLM-LABEL: @umul_v64i8(1334; SLM-NEXT:    [[TMP1:%.*]] = load <16 x i8>, ptr @a8, align 11335; SLM-NEXT:    [[TMP2:%.*]] = load <16 x i8>, ptr @b8, align 11336; SLM-NEXT:    [[TMP3:%.*]] = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> [[TMP1]], <16 x i8> [[TMP2]], i32 3)1337; SLM-NEXT:    store <16 x i8> [[TMP3]], ptr @c8, align 11338; SLM-NEXT:    [[TMP4:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 11339; SLM-NEXT:    [[TMP5:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 11340; SLM-NEXT:    [[TMP6:%.*]] = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> [[TMP4]], <16 x i8> [[TMP5]], i32 3)1341; SLM-NEXT:    store <16 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 11342; SLM-NEXT:    [[TMP7:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 11343; SLM-NEXT:    [[TMP8:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 11344; SLM-NEXT:    [[TMP9:%.*]] = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> [[TMP7]], <16 x i8> [[TMP8]], i32 3)1345; SLM-NEXT:    store <16 x i8> [[TMP9]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 11346; SLM-NEXT:    [[TMP10:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 11347; SLM-NEXT:    [[TMP11:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 11348; SLM-NEXT:    [[TMP12:%.*]] = call <16 x i8> @llvm.umul.fix.v16i8(<16 x i8> [[TMP10]], <16 x i8> [[TMP11]], i32 3)1349; SLM-NEXT:    store <16 x i8> [[TMP12]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 11350; SLM-NEXT:    ret void1351;1352; AVX-LABEL: @umul_v64i8(1353; AVX-NEXT:    [[TMP1:%.*]] = load <32 x i8>, ptr @a8, align 11354; AVX-NEXT:    [[TMP2:%.*]] = load <32 x i8>, ptr @b8, align 11355; AVX-NEXT:    [[TMP3:%.*]] = call <32 x i8> @llvm.umul.fix.v32i8(<32 x i8> [[TMP1]], <32 x i8> [[TMP2]], i32 3)1356; AVX-NEXT:    store <32 x i8> [[TMP3]], ptr @c8, align 11357; AVX-NEXT:    [[TMP4:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 11358; AVX-NEXT:    [[TMP5:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 11359; AVX-NEXT:    [[TMP6:%.*]] = call <32 x i8> @llvm.umul.fix.v32i8(<32 x i8> [[TMP4]], <32 x i8> [[TMP5]], i32 3)1360; AVX-NEXT:    store <32 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 11361; AVX-NEXT:    ret void1362;1363; AVX512-LABEL: @umul_v64i8(1364; AVX512-NEXT:    [[TMP1:%.*]] = load <64 x i8>, ptr @a8, align 11365; AVX512-NEXT:    [[TMP2:%.*]] = load <64 x i8>, ptr @b8, align 11366; AVX512-NEXT:    [[TMP3:%.*]] = call <64 x i8> @llvm.umul.fix.v64i8(<64 x i8> [[TMP1]], <64 x i8> [[TMP2]], i32 3)1367; AVX512-NEXT:    store <64 x i8> [[TMP3]], ptr @c8, align 11368; AVX512-NEXT:    ret void1369;1370  %a0  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 0 ), align 11371  %a1  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 1 ), align 11372  %a2  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 2 ), align 11373  %a3  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 3 ), align 11374  %a4  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 4 ), align 11375  %a5  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 5 ), align 11376  %a6  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 6 ), align 11377  %a7  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 7 ), align 11378  %a8  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 8 ), align 11379  %a9  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 9 ), align 11380  %a10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 10), align 11381  %a11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 11), align 11382  %a12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 12), align 11383  %a13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 13), align 11384  %a14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 14), align 11385  %a15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 15), align 11386  %a16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 11387  %a17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 17), align 11388  %a18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 18), align 11389  %a19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 19), align 11390  %a20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 20), align 11391  %a21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 21), align 11392  %a22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 22), align 11393  %a23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 23), align 11394  %a24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 24), align 11395  %a25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 25), align 11396  %a26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 26), align 11397  %a27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 27), align 11398  %a28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 28), align 11399  %a29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 29), align 11400  %a30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 30), align 11401  %a31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 31), align 11402  %a32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 11403  %a33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 33), align 11404  %a34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 34), align 11405  %a35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 35), align 11406  %a36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 36), align 11407  %a37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 37), align 11408  %a38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 38), align 11409  %a39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 39), align 11410  %a40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 40), align 11411  %a41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 41), align 11412  %a42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 42), align 11413  %a43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 43), align 11414  %a44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 44), align 11415  %a45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 45), align 11416  %a46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 46), align 11417  %a47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 47), align 11418  %a48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 11419  %a49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 49), align 11420  %a50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 50), align 11421  %a51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 51), align 11422  %a52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 52), align 11423  %a53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 53), align 11424  %a54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 54), align 11425  %a55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 55), align 11426  %a56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 56), align 11427  %a57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 57), align 11428  %a58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 58), align 11429  %a59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 59), align 11430  %a60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 60), align 11431  %a61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 61), align 11432  %a62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 62), align 11433  %a63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 63), align 11434  %b0  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 0 ), align 11435  %b1  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 1 ), align 11436  %b2  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 2 ), align 11437  %b3  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 3 ), align 11438  %b4  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 4 ), align 11439  %b5  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 5 ), align 11440  %b6  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 6 ), align 11441  %b7  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 7 ), align 11442  %b8  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 8 ), align 11443  %b9  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 9 ), align 11444  %b10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 10), align 11445  %b11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 11), align 11446  %b12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 12), align 11447  %b13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 13), align 11448  %b14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 14), align 11449  %b15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 15), align 11450  %b16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 11451  %b17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 17), align 11452  %b18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 18), align 11453  %b19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 19), align 11454  %b20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 20), align 11455  %b21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 21), align 11456  %b22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 22), align 11457  %b23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 23), align 11458  %b24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 24), align 11459  %b25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 25), align 11460  %b26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 26), align 11461  %b27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 27), align 11462  %b28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 28), align 11463  %b29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 29), align 11464  %b30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 30), align 11465  %b31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 31), align 11466  %b32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 11467  %b33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 33), align 11468  %b34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 34), align 11469  %b35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 35), align 11470  %b36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 36), align 11471  %b37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 37), align 11472  %b38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 38), align 11473  %b39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 39), align 11474  %b40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 40), align 11475  %b41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 41), align 11476  %b42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 42), align 11477  %b43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 43), align 11478  %b44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 44), align 11479  %b45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 45), align 11480  %b46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 46), align 11481  %b47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 47), align 11482  %b48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 11483  %b49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 49), align 11484  %b50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 50), align 11485  %b51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 51), align 11486  %b52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 52), align 11487  %b53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 53), align 11488  %b54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 54), align 11489  %b55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 55), align 11490  %b56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 56), align 11491  %b57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 57), align 11492  %b58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 58), align 11493  %b59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 59), align 11494  %b60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 60), align 11495  %b61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 61), align 11496  %b62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 62), align 11497  %b63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 63), align 11498  %r0  = call i8 @llvm.umul.fix.i8(i8 %a0 , i8 %b0 , i32 3)1499  %r1  = call i8 @llvm.umul.fix.i8(i8 %a1 , i8 %b1 , i32 3)1500  %r2  = call i8 @llvm.umul.fix.i8(i8 %a2 , i8 %b2 , i32 3)1501  %r3  = call i8 @llvm.umul.fix.i8(i8 %a3 , i8 %b3 , i32 3)1502  %r4  = call i8 @llvm.umul.fix.i8(i8 %a4 , i8 %b4 , i32 3)1503  %r5  = call i8 @llvm.umul.fix.i8(i8 %a5 , i8 %b5 , i32 3)1504  %r6  = call i8 @llvm.umul.fix.i8(i8 %a6 , i8 %b6 , i32 3)1505  %r7  = call i8 @llvm.umul.fix.i8(i8 %a7 , i8 %b7 , i32 3)1506  %r8  = call i8 @llvm.umul.fix.i8(i8 %a8 , i8 %b8 , i32 3)1507  %r9  = call i8 @llvm.umul.fix.i8(i8 %a9 , i8 %b9 , i32 3)1508  %r10 = call i8 @llvm.umul.fix.i8(i8 %a10, i8 %b10, i32 3)1509  %r11 = call i8 @llvm.umul.fix.i8(i8 %a11, i8 %b11, i32 3)1510  %r12 = call i8 @llvm.umul.fix.i8(i8 %a12, i8 %b12, i32 3)1511  %r13 = call i8 @llvm.umul.fix.i8(i8 %a13, i8 %b13, i32 3)1512  %r14 = call i8 @llvm.umul.fix.i8(i8 %a14, i8 %b14, i32 3)1513  %r15 = call i8 @llvm.umul.fix.i8(i8 %a15, i8 %b15, i32 3)1514  %r16 = call i8 @llvm.umul.fix.i8(i8 %a16, i8 %b16, i32 3)1515  %r17 = call i8 @llvm.umul.fix.i8(i8 %a17, i8 %b17, i32 3)1516  %r18 = call i8 @llvm.umul.fix.i8(i8 %a18, i8 %b18, i32 3)1517  %r19 = call i8 @llvm.umul.fix.i8(i8 %a19, i8 %b19, i32 3)1518  %r20 = call i8 @llvm.umul.fix.i8(i8 %a20, i8 %b20, i32 3)1519  %r21 = call i8 @llvm.umul.fix.i8(i8 %a21, i8 %b21, i32 3)1520  %r22 = call i8 @llvm.umul.fix.i8(i8 %a22, i8 %b22, i32 3)1521  %r23 = call i8 @llvm.umul.fix.i8(i8 %a23, i8 %b23, i32 3)1522  %r24 = call i8 @llvm.umul.fix.i8(i8 %a24, i8 %b24, i32 3)1523  %r25 = call i8 @llvm.umul.fix.i8(i8 %a25, i8 %b25, i32 3)1524  %r26 = call i8 @llvm.umul.fix.i8(i8 %a26, i8 %b26, i32 3)1525  %r27 = call i8 @llvm.umul.fix.i8(i8 %a27, i8 %b27, i32 3)1526  %r28 = call i8 @llvm.umul.fix.i8(i8 %a28, i8 %b28, i32 3)1527  %r29 = call i8 @llvm.umul.fix.i8(i8 %a29, i8 %b29, i32 3)1528  %r30 = call i8 @llvm.umul.fix.i8(i8 %a30, i8 %b30, i32 3)1529  %r31 = call i8 @llvm.umul.fix.i8(i8 %a31, i8 %b31, i32 3)1530  %r32 = call i8 @llvm.umul.fix.i8(i8 %a32, i8 %b32, i32 3)1531  %r33 = call i8 @llvm.umul.fix.i8(i8 %a33, i8 %b33, i32 3)1532  %r34 = call i8 @llvm.umul.fix.i8(i8 %a34, i8 %b34, i32 3)1533  %r35 = call i8 @llvm.umul.fix.i8(i8 %a35, i8 %b35, i32 3)1534  %r36 = call i8 @llvm.umul.fix.i8(i8 %a36, i8 %b36, i32 3)1535  %r37 = call i8 @llvm.umul.fix.i8(i8 %a37, i8 %b37, i32 3)1536  %r38 = call i8 @llvm.umul.fix.i8(i8 %a38, i8 %b38, i32 3)1537  %r39 = call i8 @llvm.umul.fix.i8(i8 %a39, i8 %b39, i32 3)1538  %r40 = call i8 @llvm.umul.fix.i8(i8 %a40, i8 %b40, i32 3)1539  %r41 = call i8 @llvm.umul.fix.i8(i8 %a41, i8 %b41, i32 3)1540  %r42 = call i8 @llvm.umul.fix.i8(i8 %a42, i8 %b42, i32 3)1541  %r43 = call i8 @llvm.umul.fix.i8(i8 %a43, i8 %b43, i32 3)1542  %r44 = call i8 @llvm.umul.fix.i8(i8 %a44, i8 %b44, i32 3)1543  %r45 = call i8 @llvm.umul.fix.i8(i8 %a45, i8 %b45, i32 3)1544  %r46 = call i8 @llvm.umul.fix.i8(i8 %a46, i8 %b46, i32 3)1545  %r47 = call i8 @llvm.umul.fix.i8(i8 %a47, i8 %b47, i32 3)1546  %r48 = call i8 @llvm.umul.fix.i8(i8 %a48, i8 %b48, i32 3)1547  %r49 = call i8 @llvm.umul.fix.i8(i8 %a49, i8 %b49, i32 3)1548  %r50 = call i8 @llvm.umul.fix.i8(i8 %a50, i8 %b50, i32 3)1549  %r51 = call i8 @llvm.umul.fix.i8(i8 %a51, i8 %b51, i32 3)1550  %r52 = call i8 @llvm.umul.fix.i8(i8 %a52, i8 %b52, i32 3)1551  %r53 = call i8 @llvm.umul.fix.i8(i8 %a53, i8 %b53, i32 3)1552  %r54 = call i8 @llvm.umul.fix.i8(i8 %a54, i8 %b54, i32 3)1553  %r55 = call i8 @llvm.umul.fix.i8(i8 %a55, i8 %b55, i32 3)1554  %r56 = call i8 @llvm.umul.fix.i8(i8 %a56, i8 %b56, i32 3)1555  %r57 = call i8 @llvm.umul.fix.i8(i8 %a57, i8 %b57, i32 3)1556  %r58 = call i8 @llvm.umul.fix.i8(i8 %a58, i8 %b58, i32 3)1557  %r59 = call i8 @llvm.umul.fix.i8(i8 %a59, i8 %b59, i32 3)1558  %r60 = call i8 @llvm.umul.fix.i8(i8 %a60, i8 %b60, i32 3)1559  %r61 = call i8 @llvm.umul.fix.i8(i8 %a61, i8 %b61, i32 3)1560  %r62 = call i8 @llvm.umul.fix.i8(i8 %a62, i8 %b62, i32 3)1561  %r63 = call i8 @llvm.umul.fix.i8(i8 %a63, i8 %b63, i32 3)1562  store i8 %r0 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 0 ), align 11563  store i8 %r1 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 1 ), align 11564  store i8 %r2 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 2 ), align 11565  store i8 %r3 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 3 ), align 11566  store i8 %r4 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 4 ), align 11567  store i8 %r5 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 5 ), align 11568  store i8 %r6 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 6 ), align 11569  store i8 %r7 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 7 ), align 11570  store i8 %r8 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 8 ), align 11571  store i8 %r9 , ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 9 ), align 11572  store i8 %r10, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 10), align 11573  store i8 %r11, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 11), align 11574  store i8 %r12, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 12), align 11575  store i8 %r13, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 13), align 11576  store i8 %r14, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 14), align 11577  store i8 %r15, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 15), align 11578  store i8 %r16, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 16), align 11579  store i8 %r17, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 17), align 11580  store i8 %r18, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 18), align 11581  store i8 %r19, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 19), align 11582  store i8 %r20, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 20), align 11583  store i8 %r21, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 21), align 11584  store i8 %r22, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 22), align 11585  store i8 %r23, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 23), align 11586  store i8 %r24, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 24), align 11587  store i8 %r25, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 25), align 11588  store i8 %r26, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 26), align 11589  store i8 %r27, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 27), align 11590  store i8 %r28, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 28), align 11591  store i8 %r29, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 29), align 11592  store i8 %r30, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 30), align 11593  store i8 %r31, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 31), align 11594  store i8 %r32, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 32), align 11595  store i8 %r33, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 33), align 11596  store i8 %r34, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 34), align 11597  store i8 %r35, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 35), align 11598  store i8 %r36, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 36), align 11599  store i8 %r37, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 37), align 11600  store i8 %r38, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 38), align 11601  store i8 %r39, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 39), align 11602  store i8 %r40, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 40), align 11603  store i8 %r41, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 41), align 11604  store i8 %r42, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 42), align 11605  store i8 %r43, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 43), align 11606  store i8 %r44, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 44), align 11607  store i8 %r45, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 45), align 11608  store i8 %r46, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 46), align 11609  store i8 %r47, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 47), align 11610  store i8 %r48, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 48), align 11611  store i8 %r49, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 49), align 11612  store i8 %r50, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 50), align 11613  store i8 %r51, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 51), align 11614  store i8 %r52, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 52), align 11615  store i8 %r53, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 53), align 11616  store i8 %r54, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 54), align 11617  store i8 %r55, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 55), align 11618  store i8 %r56, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 56), align 11619  store i8 %r57, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 57), align 11620  store i8 %r58, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 58), align 11621  store i8 %r59, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 59), align 11622  store i8 %r60, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 60), align 11623  store i8 %r61, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 61), align 11624  store i8 %r62, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 62), align 11625  store i8 %r63, ptr getelementptr inbounds ([64 x i8], ptr @c8, i32 0, i64 63), align 11626  ret void1627}1628