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=CHECK,SSE3; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=slm -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=CHECK,SSE4; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=corei7-avx -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=CHECK,AVX,AVX15; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=core-avx2 -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=CHECK,AVX,AVX26; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=skx -mattr=+prefer-256-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=CHECK,AVX,AVX2567; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=skx -mattr=-prefer-256-bit -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=CHECK,AVX5128; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=znver4 -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=AVX512VBMI29; RUN: opt < %s -mtriple=x86_64-unknown -mcpu=knl -passes=slp-vectorizer -S | FileCheck %s --check-prefixes=CHECK,AVX51210 11@a64 = common global [8 x i64] zeroinitializer, align 6412@b64 = common global [8 x i64] zeroinitializer, align 6413@c64 = common global [8 x i64] zeroinitializer, align 6414@d64 = common global [8 x i64] zeroinitializer, align 6415@a32 = common global [16 x i32] zeroinitializer, align 6416@b32 = common global [16 x i32] zeroinitializer, align 6417@c32 = common global [16 x i32] zeroinitializer, align 6418@d32 = common global [16 x i32] zeroinitializer, align 6419@a16 = common global [32 x i16] zeroinitializer, align 6420@b16 = common global [32 x i16] zeroinitializer, align 6421@c16 = common global [32 x i16] zeroinitializer, align 6422@d16 = common global [32 x i16] zeroinitializer, align 6423@a8  = common global [64 x i8] zeroinitializer, align 6424@b8  = common global [64 x i8] zeroinitializer, align 6425@c8  = common global [64 x i8] zeroinitializer, align 6426@d8  = common global [64 x i8] zeroinitializer, align 6427 28declare i64 @llvm.fshr.i64(i64, i64, i64)29declare i32 @llvm.fshr.i32(i32, i32, i32)30declare i16 @llvm.fshr.i16(i16, i16, i16)31declare i8  @llvm.fshr.i8 (i8 , i8 , i8 )32 33define void @fshr_v8i64() {34; SSE-LABEL: @fshr_v8i64(35; SSE-NEXT:    [[A0:%.*]] = load i64, ptr @a64, align 836; SSE-NEXT:    [[A1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 837; SSE-NEXT:    [[A2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 838; SSE-NEXT:    [[A3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 839; SSE-NEXT:    [[A4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 840; SSE-NEXT:    [[A5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 841; SSE-NEXT:    [[A6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 842; SSE-NEXT:    [[A7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 843; SSE-NEXT:    [[B0:%.*]] = load i64, ptr @b64, align 844; SSE-NEXT:    [[B1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 845; SSE-NEXT:    [[B2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 846; SSE-NEXT:    [[B3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 847; SSE-NEXT:    [[B4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 848; SSE-NEXT:    [[B5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 849; SSE-NEXT:    [[B6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 850; SSE-NEXT:    [[B7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 851; SSE-NEXT:    [[R0:%.*]] = call i64 @llvm.fshr.i64(i64 [[A0]], i64 [[A0]], i64 [[B0]])52; SSE-NEXT:    [[R1:%.*]] = call i64 @llvm.fshr.i64(i64 [[A1]], i64 [[A1]], i64 [[B1]])53; SSE-NEXT:    [[R2:%.*]] = call i64 @llvm.fshr.i64(i64 [[A2]], i64 [[A2]], i64 [[B2]])54; SSE-NEXT:    [[R3:%.*]] = call i64 @llvm.fshr.i64(i64 [[A3]], i64 [[A3]], i64 [[B3]])55; SSE-NEXT:    [[R4:%.*]] = call i64 @llvm.fshr.i64(i64 [[A4]], i64 [[A4]], i64 [[B4]])56; SSE-NEXT:    [[R5:%.*]] = call i64 @llvm.fshr.i64(i64 [[A5]], i64 [[A5]], i64 [[B5]])57; SSE-NEXT:    [[R6:%.*]] = call i64 @llvm.fshr.i64(i64 [[A6]], i64 [[A6]], i64 [[B6]])58; SSE-NEXT:    [[R7:%.*]] = call i64 @llvm.fshr.i64(i64 [[A7]], i64 [[A7]], i64 [[B7]])59; SSE-NEXT:    store i64 [[R0]], ptr @d64, align 860; SSE-NEXT:    store i64 [[R1]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 1), align 861; SSE-NEXT:    store i64 [[R2]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 2), align 862; SSE-NEXT:    store i64 [[R3]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 3), align 863; SSE-NEXT:    store i64 [[R4]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 864; SSE-NEXT:    store i64 [[R5]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 5), align 865; SSE-NEXT:    store i64 [[R6]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 6), align 866; SSE-NEXT:    store i64 [[R7]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 7), align 867; SSE-NEXT:    ret void68;69; AVX1-LABEL: @fshr_v8i64(70; AVX1-NEXT:    [[A0:%.*]] = load i64, ptr @a64, align 871; AVX1-NEXT:    [[A1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 872; AVX1-NEXT:    [[A2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 873; AVX1-NEXT:    [[A3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 874; AVX1-NEXT:    [[A4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 875; AVX1-NEXT:    [[A5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 876; AVX1-NEXT:    [[A6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 877; AVX1-NEXT:    [[A7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 878; AVX1-NEXT:    [[B0:%.*]] = load i64, ptr @b64, align 879; AVX1-NEXT:    [[B1:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 880; AVX1-NEXT:    [[B2:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 881; AVX1-NEXT:    [[B3:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 882; AVX1-NEXT:    [[B4:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 883; AVX1-NEXT:    [[B5:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 884; AVX1-NEXT:    [[B6:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 885; AVX1-NEXT:    [[B7:%.*]] = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 886; AVX1-NEXT:    [[R0:%.*]] = call i64 @llvm.fshr.i64(i64 [[A0]], i64 [[A0]], i64 [[B0]])87; AVX1-NEXT:    [[R1:%.*]] = call i64 @llvm.fshr.i64(i64 [[A1]], i64 [[A1]], i64 [[B1]])88; AVX1-NEXT:    [[R2:%.*]] = call i64 @llvm.fshr.i64(i64 [[A2]], i64 [[A2]], i64 [[B2]])89; AVX1-NEXT:    [[R3:%.*]] = call i64 @llvm.fshr.i64(i64 [[A3]], i64 [[A3]], i64 [[B3]])90; AVX1-NEXT:    [[R4:%.*]] = call i64 @llvm.fshr.i64(i64 [[A4]], i64 [[A4]], i64 [[B4]])91; AVX1-NEXT:    [[R5:%.*]] = call i64 @llvm.fshr.i64(i64 [[A5]], i64 [[A5]], i64 [[B5]])92; AVX1-NEXT:    [[R6:%.*]] = call i64 @llvm.fshr.i64(i64 [[A6]], i64 [[A6]], i64 [[B6]])93; AVX1-NEXT:    [[R7:%.*]] = call i64 @llvm.fshr.i64(i64 [[A7]], i64 [[A7]], i64 [[B7]])94; AVX1-NEXT:    store i64 [[R0]], ptr @d64, align 895; AVX1-NEXT:    store i64 [[R1]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 1), align 896; AVX1-NEXT:    store i64 [[R2]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 2), align 897; AVX1-NEXT:    store i64 [[R3]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 3), align 898; AVX1-NEXT:    store i64 [[R4]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 899; AVX1-NEXT:    store i64 [[R5]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 5), align 8100; AVX1-NEXT:    store i64 [[R6]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 6), align 8101; AVX1-NEXT:    store i64 [[R7]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 7), align 8102; AVX1-NEXT:    ret void103;104; AVX2-LABEL: @fshr_v8i64(105; AVX2-NEXT:    [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 8106; AVX2-NEXT:    [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 8107; AVX2-NEXT:    [[TMP3:%.*]] = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> [[TMP1]], <4 x i64> [[TMP1]], <4 x i64> [[TMP2]])108; AVX2-NEXT:    store <4 x i64> [[TMP3]], ptr @d64, align 8109; AVX2-NEXT:    [[TMP4:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8110; AVX2-NEXT:    [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8111; AVX2-NEXT:    [[TMP6:%.*]] = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> [[TMP4]], <4 x i64> [[TMP4]], <4 x i64> [[TMP5]])112; AVX2-NEXT:    store <4 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 8113; AVX2-NEXT:    ret void114;115; AVX256-LABEL: @fshr_v8i64(116; AVX256-NEXT:    [[TMP1:%.*]] = load <4 x i64>, ptr @a64, align 8117; AVX256-NEXT:    [[TMP2:%.*]] = load <4 x i64>, ptr @b64, align 8118; AVX256-NEXT:    [[TMP3:%.*]] = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> [[TMP1]], <4 x i64> [[TMP1]], <4 x i64> [[TMP2]])119; AVX256-NEXT:    store <4 x i64> [[TMP3]], ptr @d64, align 8120; AVX256-NEXT:    [[TMP4:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8121; AVX256-NEXT:    [[TMP5:%.*]] = load <4 x i64>, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8122; AVX256-NEXT:    [[TMP6:%.*]] = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> [[TMP4]], <4 x i64> [[TMP4]], <4 x i64> [[TMP5]])123; AVX256-NEXT:    store <4 x i64> [[TMP6]], ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 8124; AVX256-NEXT:    ret void125;126; AVX512-LABEL: @fshr_v8i64(127; AVX512-NEXT:    [[TMP1:%.*]] = load <8 x i64>, ptr @a64, align 8128; AVX512-NEXT:    [[TMP2:%.*]] = load <8 x i64>, ptr @b64, align 8129; AVX512-NEXT:    [[TMP3:%.*]] = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> [[TMP1]], <8 x i64> [[TMP1]], <8 x i64> [[TMP2]])130; AVX512-NEXT:    store <8 x i64> [[TMP3]], ptr @d64, align 8131; AVX512-NEXT:    ret void132;133; AVX512VBMI2-LABEL: @fshr_v8i64(134; AVX512VBMI2-NEXT:    [[TMP1:%.*]] = load <8 x i64>, ptr @a64, align 8135; AVX512VBMI2-NEXT:    [[TMP2:%.*]] = load <8 x i64>, ptr @b64, align 8136; AVX512VBMI2-NEXT:    [[TMP3:%.*]] = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> [[TMP1]], <8 x i64> [[TMP1]], <8 x i64> [[TMP2]])137; AVX512VBMI2-NEXT:    store <8 x i64> [[TMP3]], ptr @d64, align 8138; AVX512VBMI2-NEXT:    ret void139;140  %a0 = load i64, ptr @a64, align 8141  %a1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 1), align 8142  %a2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 2), align 8143  %a3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 3), align 8144  %a4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 4), align 8145  %a5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 5), align 8146  %a6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 6), align 8147  %a7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @a64, i32 0, i64 7), align 8148  %b0 = load i64, ptr @b64, align 8149  %b1 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 1), align 8150  %b2 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 2), align 8151  %b3 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 3), align 8152  %b4 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 4), align 8153  %b5 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 5), align 8154  %b6 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 6), align 8155  %b7 = load i64, ptr getelementptr inbounds ([8 x i64], ptr @b64, i32 0, i64 7), align 8156  %r0 = call i64 @llvm.fshr.i64(i64 %a0, i64 %a0, i64 %b0)157  %r1 = call i64 @llvm.fshr.i64(i64 %a1, i64 %a1, i64 %b1)158  %r2 = call i64 @llvm.fshr.i64(i64 %a2, i64 %a2, i64 %b2)159  %r3 = call i64 @llvm.fshr.i64(i64 %a3, i64 %a3, i64 %b3)160  %r4 = call i64 @llvm.fshr.i64(i64 %a4, i64 %a4, i64 %b4)161  %r5 = call i64 @llvm.fshr.i64(i64 %a5, i64 %a5, i64 %b5)162  %r6 = call i64 @llvm.fshr.i64(i64 %a6, i64 %a6, i64 %b6)163  %r7 = call i64 @llvm.fshr.i64(i64 %a7, i64 %a7, i64 %b7)164  store i64 %r0, ptr @d64, align 8165  store i64 %r1, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 1), align 8166  store i64 %r2, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 2), align 8167  store i64 %r3, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 3), align 8168  store i64 %r4, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 4), align 8169  store i64 %r5, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 5), align 8170  store i64 %r6, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 6), align 8171  store i64 %r7, ptr getelementptr inbounds ([8 x i64], ptr @d64, i32 0, i64 7), align 8172  ret void173}174 175define void @fshr_v16i32() {176; SSE-LABEL: @fshr_v16i32(177; SSE-NEXT:    [[A0:%.*]] = load i32, ptr @a32, align 4178; SSE-NEXT:    [[A1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1), align 4179; SSE-NEXT:    [[A2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2), align 4180; SSE-NEXT:    [[A3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3), align 4181; SSE-NEXT:    [[A4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4182; SSE-NEXT:    [[A5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5), align 4183; SSE-NEXT:    [[A6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6), align 4184; SSE-NEXT:    [[A7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7), align 4185; SSE-NEXT:    [[A8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4186; SSE-NEXT:    [[A9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9), align 4187; SSE-NEXT:    [[A10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 4188; SSE-NEXT:    [[A11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 4189; SSE-NEXT:    [[A12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4190; SSE-NEXT:    [[A13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 4191; SSE-NEXT:    [[A14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 4192; SSE-NEXT:    [[A15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 4193; SSE-NEXT:    [[B0:%.*]] = load i32, ptr @b32, align 4194; SSE-NEXT:    [[B1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1), align 4195; SSE-NEXT:    [[B2:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2), align 4196; SSE-NEXT:    [[B3:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3), align 4197; SSE-NEXT:    [[B4:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4198; SSE-NEXT:    [[B5:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5), align 4199; SSE-NEXT:    [[B6:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6), align 4200; SSE-NEXT:    [[B7:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7), align 4201; SSE-NEXT:    [[B8:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4202; SSE-NEXT:    [[B9:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9), align 4203; SSE-NEXT:    [[B10:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 4204; SSE-NEXT:    [[B11:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 4205; SSE-NEXT:    [[B12:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4206; SSE-NEXT:    [[B13:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 4207; SSE-NEXT:    [[B14:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 4208; SSE-NEXT:    [[B15:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 4209; SSE-NEXT:    [[R0:%.*]] = call i32 @llvm.fshr.i32(i32 [[A0]], i32 [[A0]], i32 [[B0]])210; SSE-NEXT:    [[R1:%.*]] = call i32 @llvm.fshr.i32(i32 [[A1]], i32 [[A1]], i32 [[B1]])211; SSE-NEXT:    [[R2:%.*]] = call i32 @llvm.fshr.i32(i32 [[A2]], i32 [[A2]], i32 [[B2]])212; SSE-NEXT:    [[R3:%.*]] = call i32 @llvm.fshr.i32(i32 [[A3]], i32 [[A3]], i32 [[B3]])213; SSE-NEXT:    [[R4:%.*]] = call i32 @llvm.fshr.i32(i32 [[A4]], i32 [[A4]], i32 [[B4]])214; SSE-NEXT:    [[R5:%.*]] = call i32 @llvm.fshr.i32(i32 [[A5]], i32 [[A5]], i32 [[B5]])215; SSE-NEXT:    [[R6:%.*]] = call i32 @llvm.fshr.i32(i32 [[A6]], i32 [[A6]], i32 [[B6]])216; SSE-NEXT:    [[R7:%.*]] = call i32 @llvm.fshr.i32(i32 [[A7]], i32 [[A7]], i32 [[B7]])217; SSE-NEXT:    [[R8:%.*]] = call i32 @llvm.fshr.i32(i32 [[A8]], i32 [[A8]], i32 [[B8]])218; SSE-NEXT:    [[R9:%.*]] = call i32 @llvm.fshr.i32(i32 [[A9]], i32 [[A9]], i32 [[B9]])219; SSE-NEXT:    [[R10:%.*]] = call i32 @llvm.fshr.i32(i32 [[A10]], i32 [[A10]], i32 [[B10]])220; SSE-NEXT:    [[R11:%.*]] = call i32 @llvm.fshr.i32(i32 [[A11]], i32 [[A11]], i32 [[B11]])221; SSE-NEXT:    [[R12:%.*]] = call i32 @llvm.fshr.i32(i32 [[A12]], i32 [[A12]], i32 [[B12]])222; SSE-NEXT:    [[R13:%.*]] = call i32 @llvm.fshr.i32(i32 [[A13]], i32 [[A13]], i32 [[B13]])223; SSE-NEXT:    [[R14:%.*]] = call i32 @llvm.fshr.i32(i32 [[A14]], i32 [[A14]], i32 [[B14]])224; SSE-NEXT:    [[R15:%.*]] = call i32 @llvm.fshr.i32(i32 [[A15]], i32 [[A15]], i32 [[B15]])225; SSE-NEXT:    store i32 [[R0]], ptr @d32, align 4226; SSE-NEXT:    store i32 [[R1]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 1), align 4227; SSE-NEXT:    store i32 [[R2]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 2), align 4228; SSE-NEXT:    store i32 [[R3]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 3), align 4229; SSE-NEXT:    store i32 [[R4]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 4), align 4230; SSE-NEXT:    store i32 [[R5]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 5), align 4231; SSE-NEXT:    store i32 [[R6]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 6), align 4232; SSE-NEXT:    store i32 [[R7]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 7), align 4233; SSE-NEXT:    store i32 [[R8]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 8), align 4234; SSE-NEXT:    store i32 [[R9]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 9), align 4235; SSE-NEXT:    store i32 [[R10]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 10), align 4236; SSE-NEXT:    store i32 [[R11]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 11), align 4237; SSE-NEXT:    store i32 [[R12]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 12), align 4238; SSE-NEXT:    store i32 [[R13]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 13), align 4239; SSE-NEXT:    store i32 [[R14]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 14), align 4240; SSE-NEXT:    store i32 [[R15]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 15), align 4241; SSE-NEXT:    ret void242;243; AVX1-LABEL: @fshr_v16i32(244; AVX1-NEXT:    [[TMP1:%.*]] = load <4 x i32>, ptr @a32, align 4245; AVX1-NEXT:    [[TMP2:%.*]] = load <4 x i32>, ptr @b32, align 4246; AVX1-NEXT:    [[TMP3:%.*]] = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> [[TMP1]], <4 x i32> [[TMP1]], <4 x i32> [[TMP2]])247; AVX1-NEXT:    store <4 x i32> [[TMP3]], ptr @d32, align 4248; AVX1-NEXT:    [[TMP4:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4), align 4249; AVX1-NEXT:    [[TMP5:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4), align 4250; AVX1-NEXT:    [[TMP6:%.*]] = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> [[TMP4]], <4 x i32> [[TMP4]], <4 x i32> [[TMP5]])251; AVX1-NEXT:    store <4 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 4), align 4252; AVX1-NEXT:    [[TMP7:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4253; AVX1-NEXT:    [[TMP8:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4254; AVX1-NEXT:    [[TMP9:%.*]] = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> [[TMP7]], <4 x i32> [[TMP7]], <4 x i32> [[TMP8]])255; AVX1-NEXT:    store <4 x i32> [[TMP9]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 8), align 4256; AVX1-NEXT:    [[TMP10:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4257; AVX1-NEXT:    [[TMP11:%.*]] = load <4 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4258; AVX1-NEXT:    [[TMP12:%.*]] = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> [[TMP10]], <4 x i32> [[TMP10]], <4 x i32> [[TMP11]])259; AVX1-NEXT:    store <4 x i32> [[TMP12]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 12), align 4260; AVX1-NEXT:    ret void261;262; AVX2-LABEL: @fshr_v16i32(263; AVX2-NEXT:    [[TMP1:%.*]] = load <8 x i32>, ptr @a32, align 4264; AVX2-NEXT:    [[TMP2:%.*]] = load <8 x i32>, ptr @b32, align 4265; AVX2-NEXT:    [[TMP3:%.*]] = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> [[TMP1]], <8 x i32> [[TMP1]], <8 x i32> [[TMP2]])266; AVX2-NEXT:    store <8 x i32> [[TMP3]], ptr @d32, align 4267; AVX2-NEXT:    [[TMP4:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4268; AVX2-NEXT:    [[TMP5:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4269; AVX2-NEXT:    [[TMP6:%.*]] = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> [[TMP4]], <8 x i32> [[TMP4]], <8 x i32> [[TMP5]])270; AVX2-NEXT:    store <8 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 8), align 4271; AVX2-NEXT:    ret void272;273; AVX256-LABEL: @fshr_v16i32(274; AVX256-NEXT:    [[TMP1:%.*]] = load <8 x i32>, ptr @a32, align 4275; AVX256-NEXT:    [[TMP2:%.*]] = load <8 x i32>, ptr @b32, align 4276; AVX256-NEXT:    [[TMP3:%.*]] = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> [[TMP1]], <8 x i32> [[TMP1]], <8 x i32> [[TMP2]])277; AVX256-NEXT:    store <8 x i32> [[TMP3]], ptr @d32, align 4278; AVX256-NEXT:    [[TMP4:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8), align 4279; AVX256-NEXT:    [[TMP5:%.*]] = load <8 x i32>, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8), align 4280; AVX256-NEXT:    [[TMP6:%.*]] = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> [[TMP4]], <8 x i32> [[TMP4]], <8 x i32> [[TMP5]])281; AVX256-NEXT:    store <8 x i32> [[TMP6]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 8), align 4282; AVX256-NEXT:    ret void283;284; AVX512-LABEL: @fshr_v16i32(285; AVX512-NEXT:    [[TMP1:%.*]] = load <16 x i32>, ptr @a32, align 4286; AVX512-NEXT:    [[TMP2:%.*]] = load <16 x i32>, ptr @b32, align 4287; AVX512-NEXT:    [[TMP3:%.*]] = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> [[TMP1]], <16 x i32> [[TMP1]], <16 x i32> [[TMP2]])288; AVX512-NEXT:    store <16 x i32> [[TMP3]], ptr @d32, align 4289; AVX512-NEXT:    ret void290;291; AVX512VBMI2-LABEL: @fshr_v16i32(292; AVX512VBMI2-NEXT:    [[TMP1:%.*]] = load <16 x i32>, ptr @a32, align 4293; AVX512VBMI2-NEXT:    [[TMP2:%.*]] = load <16 x i32>, ptr @b32, align 4294; AVX512VBMI2-NEXT:    [[TMP3:%.*]] = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> [[TMP1]], <16 x i32> [[TMP1]], <16 x i32> [[TMP2]])295; AVX512VBMI2-NEXT:    store <16 x i32> [[TMP3]], ptr @d32, align 4296; AVX512VBMI2-NEXT:    ret void297;298  %a0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 0 ), align 4299  %a1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1 ), align 4300  %a2  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 2 ), align 4301  %a3  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 3 ), align 4302  %a4  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 4 ), align 4303  %a5  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 5 ), align 4304  %a6  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 6 ), align 4305  %a7  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 7 ), align 4306  %a8  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 8 ), align 4307  %a9  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 9 ), align 4308  %a10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 10), align 4309  %a11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 11), align 4310  %a12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 12), align 4311  %a13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 13), align 4312  %a14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 14), align 4313  %a15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 15), align 4314  %b0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 0 ), align 4315  %b1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1 ), align 4316  %b2  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 2 ), align 4317  %b3  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 3 ), align 4318  %b4  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 4 ), align 4319  %b5  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 5 ), align 4320  %b6  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 6 ), align 4321  %b7  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 7 ), align 4322  %b8  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 8 ), align 4323  %b9  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 9 ), align 4324  %b10 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 10), align 4325  %b11 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 11), align 4326  %b12 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 12), align 4327  %b13 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 13), align 4328  %b14 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 14), align 4329  %b15 = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 15), align 4330  %r0  = call i32 @llvm.fshr.i32(i32 %a0 , i32 %a0 , i32 %b0 )331  %r1  = call i32 @llvm.fshr.i32(i32 %a1 , i32 %a1 , i32 %b1 )332  %r2  = call i32 @llvm.fshr.i32(i32 %a2 , i32 %a2 , i32 %b2 )333  %r3  = call i32 @llvm.fshr.i32(i32 %a3 , i32 %a3 , i32 %b3 )334  %r4  = call i32 @llvm.fshr.i32(i32 %a4 , i32 %a4 , i32 %b4 )335  %r5  = call i32 @llvm.fshr.i32(i32 %a5 , i32 %a5 , i32 %b5 )336  %r6  = call i32 @llvm.fshr.i32(i32 %a6 , i32 %a6 , i32 %b6 )337  %r7  = call i32 @llvm.fshr.i32(i32 %a7 , i32 %a7 , i32 %b7 )338  %r8  = call i32 @llvm.fshr.i32(i32 %a8 , i32 %a8 , i32 %b8 )339  %r9  = call i32 @llvm.fshr.i32(i32 %a9 , i32 %a9 , i32 %b9 )340  %r10 = call i32 @llvm.fshr.i32(i32 %a10, i32 %a10, i32 %b10)341  %r11 = call i32 @llvm.fshr.i32(i32 %a11, i32 %a11, i32 %b11)342  %r12 = call i32 @llvm.fshr.i32(i32 %a12, i32 %a12, i32 %b12)343  %r13 = call i32 @llvm.fshr.i32(i32 %a13, i32 %a13, i32 %b13)344  %r14 = call i32 @llvm.fshr.i32(i32 %a14, i32 %a14, i32 %b14)345  %r15 = call i32 @llvm.fshr.i32(i32 %a15, i32 %a15, i32 %b15)346  store i32 %r0 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 0 ), align 4347  store i32 %r1 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 1 ), align 4348  store i32 %r2 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 2 ), align 4349  store i32 %r3 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 3 ), align 4350  store i32 %r4 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 4 ), align 4351  store i32 %r5 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 5 ), align 4352  store i32 %r6 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 6 ), align 4353  store i32 %r7 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 7 ), align 4354  store i32 %r8 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 8 ), align 4355  store i32 %r9 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 9 ), align 4356  store i32 %r10, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 10), align 4357  store i32 %r11, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 11), align 4358  store i32 %r12, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 12), align 4359  store i32 %r13, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 13), align 4360  store i32 %r14, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 14), align 4361  store i32 %r15, ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 15), align 4362  ret void363}364 365define void @fshr_v32i16() {366; SSE-LABEL: @fshr_v32i16(367; SSE-NEXT:    [[A0:%.*]] = load i16, ptr @a16, align 2368; SSE-NEXT:    [[A1:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 1), align 2369; SSE-NEXT:    [[A2:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 2), align 2370; SSE-NEXT:    [[A3:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 3), align 2371; SSE-NEXT:    [[A4:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 4), align 2372; SSE-NEXT:    [[A5:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 5), align 2373; SSE-NEXT:    [[A6:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 6), align 2374; SSE-NEXT:    [[A7:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 7), align 2375; SSE-NEXT:    [[A8:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8), align 2376; SSE-NEXT:    [[A9:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 9), align 2377; SSE-NEXT:    [[A10:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 10), align 2378; SSE-NEXT:    [[A11:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 11), align 2379; SSE-NEXT:    [[A12:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 12), align 2380; SSE-NEXT:    [[A13:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 13), align 2381; SSE-NEXT:    [[A14:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 14), align 2382; SSE-NEXT:    [[A15:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 15), align 2383; SSE-NEXT:    [[A16:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2384; SSE-NEXT:    [[A17:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 17), align 2385; SSE-NEXT:    [[A18:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 18), align 2386; SSE-NEXT:    [[A19:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 19), align 2387; SSE-NEXT:    [[A20:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 20), align 2388; SSE-NEXT:    [[A21:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 21), align 2389; SSE-NEXT:    [[A22:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 22), align 2390; SSE-NEXT:    [[A23:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 23), align 2391; SSE-NEXT:    [[A24:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2392; SSE-NEXT:    [[A25:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 25), align 2393; SSE-NEXT:    [[A26:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 26), align 2394; SSE-NEXT:    [[A27:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 27), align 2395; SSE-NEXT:    [[A28:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 28), align 2396; SSE-NEXT:    [[A29:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 29), align 2397; SSE-NEXT:    [[A30:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 30), align 2398; SSE-NEXT:    [[A31:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 31), align 2399; SSE-NEXT:    [[B0:%.*]] = load i16, ptr @b16, align 2400; SSE-NEXT:    [[B1:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 1), align 2401; SSE-NEXT:    [[B2:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 2), align 2402; SSE-NEXT:    [[B3:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 3), align 2403; SSE-NEXT:    [[B4:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 4), align 2404; SSE-NEXT:    [[B5:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 5), align 2405; SSE-NEXT:    [[B6:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 6), align 2406; SSE-NEXT:    [[B7:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 7), align 2407; SSE-NEXT:    [[B8:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8), align 2408; SSE-NEXT:    [[B9:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 9), align 2409; SSE-NEXT:    [[B10:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 10), align 2410; SSE-NEXT:    [[B11:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 11), align 2411; SSE-NEXT:    [[B12:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 12), align 2412; SSE-NEXT:    [[B13:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 13), align 2413; SSE-NEXT:    [[B14:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 14), align 2414; SSE-NEXT:    [[B15:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 15), align 2415; SSE-NEXT:    [[B16:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2416; SSE-NEXT:    [[B17:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 17), align 2417; SSE-NEXT:    [[B18:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 18), align 2418; SSE-NEXT:    [[B19:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 19), align 2419; SSE-NEXT:    [[B20:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 20), align 2420; SSE-NEXT:    [[B21:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 21), align 2421; SSE-NEXT:    [[B22:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 22), align 2422; SSE-NEXT:    [[B23:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 23), align 2423; SSE-NEXT:    [[B24:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2424; SSE-NEXT:    [[B25:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 25), align 2425; SSE-NEXT:    [[B26:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 26), align 2426; SSE-NEXT:    [[B27:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 27), align 2427; SSE-NEXT:    [[B28:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 28), align 2428; SSE-NEXT:    [[B29:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 29), align 2429; SSE-NEXT:    [[B30:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 30), align 2430; SSE-NEXT:    [[B31:%.*]] = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 31), align 2431; SSE-NEXT:    [[R0:%.*]] = call i16 @llvm.fshr.i16(i16 [[A0]], i16 [[A0]], i16 [[B0]])432; SSE-NEXT:    [[R1:%.*]] = call i16 @llvm.fshr.i16(i16 [[A1]], i16 [[A1]], i16 [[B1]])433; SSE-NEXT:    [[R2:%.*]] = call i16 @llvm.fshr.i16(i16 [[A2]], i16 [[A2]], i16 [[B2]])434; SSE-NEXT:    [[R3:%.*]] = call i16 @llvm.fshr.i16(i16 [[A3]], i16 [[A3]], i16 [[B3]])435; SSE-NEXT:    [[R4:%.*]] = call i16 @llvm.fshr.i16(i16 [[A4]], i16 [[A4]], i16 [[B4]])436; SSE-NEXT:    [[R5:%.*]] = call i16 @llvm.fshr.i16(i16 [[A5]], i16 [[A5]], i16 [[B5]])437; SSE-NEXT:    [[R6:%.*]] = call i16 @llvm.fshr.i16(i16 [[A6]], i16 [[A6]], i16 [[B6]])438; SSE-NEXT:    [[R7:%.*]] = call i16 @llvm.fshr.i16(i16 [[A7]], i16 [[A7]], i16 [[B7]])439; SSE-NEXT:    [[R8:%.*]] = call i16 @llvm.fshr.i16(i16 [[A8]], i16 [[A8]], i16 [[B8]])440; SSE-NEXT:    [[R9:%.*]] = call i16 @llvm.fshr.i16(i16 [[A9]], i16 [[A9]], i16 [[B9]])441; SSE-NEXT:    [[R10:%.*]] = call i16 @llvm.fshr.i16(i16 [[A10]], i16 [[A10]], i16 [[B10]])442; SSE-NEXT:    [[R11:%.*]] = call i16 @llvm.fshr.i16(i16 [[A11]], i16 [[A11]], i16 [[B11]])443; SSE-NEXT:    [[R12:%.*]] = call i16 @llvm.fshr.i16(i16 [[A12]], i16 [[A12]], i16 [[B12]])444; SSE-NEXT:    [[R13:%.*]] = call i16 @llvm.fshr.i16(i16 [[A13]], i16 [[A13]], i16 [[B13]])445; SSE-NEXT:    [[R14:%.*]] = call i16 @llvm.fshr.i16(i16 [[A14]], i16 [[A14]], i16 [[B14]])446; SSE-NEXT:    [[R15:%.*]] = call i16 @llvm.fshr.i16(i16 [[A15]], i16 [[A15]], i16 [[B15]])447; SSE-NEXT:    [[R16:%.*]] = call i16 @llvm.fshr.i16(i16 [[A16]], i16 [[A16]], i16 [[B16]])448; SSE-NEXT:    [[R17:%.*]] = call i16 @llvm.fshr.i16(i16 [[A17]], i16 [[A17]], i16 [[B17]])449; SSE-NEXT:    [[R18:%.*]] = call i16 @llvm.fshr.i16(i16 [[A18]], i16 [[A18]], i16 [[B18]])450; SSE-NEXT:    [[R19:%.*]] = call i16 @llvm.fshr.i16(i16 [[A19]], i16 [[A19]], i16 [[B19]])451; SSE-NEXT:    [[R20:%.*]] = call i16 @llvm.fshr.i16(i16 [[A20]], i16 [[A20]], i16 [[B20]])452; SSE-NEXT:    [[R21:%.*]] = call i16 @llvm.fshr.i16(i16 [[A21]], i16 [[A21]], i16 [[B21]])453; SSE-NEXT:    [[R22:%.*]] = call i16 @llvm.fshr.i16(i16 [[A22]], i16 [[A22]], i16 [[B22]])454; SSE-NEXT:    [[R23:%.*]] = call i16 @llvm.fshr.i16(i16 [[A23]], i16 [[A23]], i16 [[B23]])455; SSE-NEXT:    [[R24:%.*]] = call i16 @llvm.fshr.i16(i16 [[A24]], i16 [[A24]], i16 [[B24]])456; SSE-NEXT:    [[R25:%.*]] = call i16 @llvm.fshr.i16(i16 [[A25]], i16 [[A25]], i16 [[B25]])457; SSE-NEXT:    [[R26:%.*]] = call i16 @llvm.fshr.i16(i16 [[A26]], i16 [[A26]], i16 [[B26]])458; SSE-NEXT:    [[R27:%.*]] = call i16 @llvm.fshr.i16(i16 [[A27]], i16 [[A27]], i16 [[B27]])459; SSE-NEXT:    [[R28:%.*]] = call i16 @llvm.fshr.i16(i16 [[A28]], i16 [[A28]], i16 [[B28]])460; SSE-NEXT:    [[R29:%.*]] = call i16 @llvm.fshr.i16(i16 [[A29]], i16 [[A29]], i16 [[B29]])461; SSE-NEXT:    [[R30:%.*]] = call i16 @llvm.fshr.i16(i16 [[A30]], i16 [[A30]], i16 [[B30]])462; SSE-NEXT:    [[R31:%.*]] = call i16 @llvm.fshr.i16(i16 [[A31]], i16 [[A31]], i16 [[B31]])463; SSE-NEXT:    store i16 [[R0]], ptr @d16, align 2464; SSE-NEXT:    store i16 [[R1]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 1), align 2465; SSE-NEXT:    store i16 [[R2]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 2), align 2466; SSE-NEXT:    store i16 [[R3]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 3), align 2467; SSE-NEXT:    store i16 [[R4]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 4), align 2468; SSE-NEXT:    store i16 [[R5]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 5), align 2469; SSE-NEXT:    store i16 [[R6]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 6), align 2470; SSE-NEXT:    store i16 [[R7]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 7), align 2471; SSE-NEXT:    store i16 [[R8]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 8), align 2472; SSE-NEXT:    store i16 [[R9]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 9), align 2473; SSE-NEXT:    store i16 [[R10]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 10), align 2474; SSE-NEXT:    store i16 [[R11]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 11), align 2475; SSE-NEXT:    store i16 [[R12]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 12), align 2476; SSE-NEXT:    store i16 [[R13]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 13), align 2477; SSE-NEXT:    store i16 [[R14]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 14), align 2478; SSE-NEXT:    store i16 [[R15]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 15), align 2479; SSE-NEXT:    store i16 [[R16]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 16), align 2480; SSE-NEXT:    store i16 [[R17]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 17), align 2481; SSE-NEXT:    store i16 [[R18]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 18), align 2482; SSE-NEXT:    store i16 [[R19]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 19), align 2483; SSE-NEXT:    store i16 [[R20]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 20), align 2484; SSE-NEXT:    store i16 [[R21]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 21), align 2485; SSE-NEXT:    store i16 [[R22]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 22), align 2486; SSE-NEXT:    store i16 [[R23]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 23), align 2487; SSE-NEXT:    store i16 [[R24]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 24), align 2488; SSE-NEXT:    store i16 [[R25]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 25), align 2489; SSE-NEXT:    store i16 [[R26]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 26), align 2490; SSE-NEXT:    store i16 [[R27]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 27), align 2491; SSE-NEXT:    store i16 [[R28]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 28), align 2492; SSE-NEXT:    store i16 [[R29]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 29), align 2493; SSE-NEXT:    store i16 [[R30]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 30), align 2494; SSE-NEXT:    store i16 [[R31]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 31), align 2495; SSE-NEXT:    ret void496;497; AVX-LABEL: @fshr_v32i16(498; AVX-NEXT:    [[TMP1:%.*]] = load <16 x i16>, ptr @a16, align 2499; AVX-NEXT:    [[TMP2:%.*]] = load <16 x i16>, ptr @b16, align 2500; AVX-NEXT:    [[TMP3:%.*]] = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> [[TMP1]], <16 x i16> [[TMP1]], <16 x i16> [[TMP2]])501; AVX-NEXT:    store <16 x i16> [[TMP3]], ptr @d16, align 2502; AVX-NEXT:    [[TMP4:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2503; AVX-NEXT:    [[TMP5:%.*]] = load <16 x i16>, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2504; AVX-NEXT:    [[TMP6:%.*]] = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> [[TMP4]], <16 x i16> [[TMP4]], <16 x i16> [[TMP5]])505; AVX-NEXT:    store <16 x i16> [[TMP6]], ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 16), align 2506; AVX-NEXT:    ret void507;508; AVX512-LABEL: @fshr_v32i16(509; AVX512-NEXT:    [[TMP1:%.*]] = load <32 x i16>, ptr @a16, align 2510; AVX512-NEXT:    [[TMP2:%.*]] = load <32 x i16>, ptr @b16, align 2511; AVX512-NEXT:    [[TMP3:%.*]] = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> [[TMP1]], <32 x i16> [[TMP1]], <32 x i16> [[TMP2]])512; AVX512-NEXT:    store <32 x i16> [[TMP3]], ptr @d16, align 2513; AVX512-NEXT:    ret void514;515; AVX512VBMI2-LABEL: @fshr_v32i16(516; AVX512VBMI2-NEXT:    [[TMP1:%.*]] = load <32 x i16>, ptr @a16, align 2517; AVX512VBMI2-NEXT:    [[TMP2:%.*]] = load <32 x i16>, ptr @b16, align 2518; AVX512VBMI2-NEXT:    [[TMP3:%.*]] = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> [[TMP1]], <32 x i16> [[TMP1]], <32 x i16> [[TMP2]])519; AVX512VBMI2-NEXT:    store <32 x i16> [[TMP3]], ptr @d16, align 2520; AVX512VBMI2-NEXT:    ret void521;522  %a0  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 0 ), align 2523  %a1  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 1 ), align 2524  %a2  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 2 ), align 2525  %a3  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 3 ), align 2526  %a4  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 4 ), align 2527  %a5  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 5 ), align 2528  %a6  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 6 ), align 2529  %a7  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 7 ), align 2530  %a8  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 8 ), align 2531  %a9  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 9 ), align 2532  %a10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 10), align 2533  %a11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 11), align 2534  %a12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 12), align 2535  %a13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 13), align 2536  %a14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 14), align 2537  %a15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 15), align 2538  %a16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 16), align 2539  %a17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 17), align 2540  %a18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 18), align 2541  %a19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 19), align 2542  %a20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 20), align 2543  %a21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 21), align 2544  %a22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 22), align 2545  %a23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 23), align 2546  %a24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 24), align 2547  %a25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 25), align 2548  %a26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 26), align 2549  %a27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 27), align 2550  %a28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 28), align 2551  %a29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 29), align 2552  %a30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 30), align 2553  %a31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @a16, i32 0, i64 31), align 2554  %b0  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 0 ), align 2555  %b1  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 1 ), align 2556  %b2  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 2 ), align 2557  %b3  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 3 ), align 2558  %b4  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 4 ), align 2559  %b5  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 5 ), align 2560  %b6  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 6 ), align 2561  %b7  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 7 ), align 2562  %b8  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 8 ), align 2563  %b9  = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 9 ), align 2564  %b10 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 10), align 2565  %b11 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 11), align 2566  %b12 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 12), align 2567  %b13 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 13), align 2568  %b14 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 14), align 2569  %b15 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 15), align 2570  %b16 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 16), align 2571  %b17 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 17), align 2572  %b18 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 18), align 2573  %b19 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 19), align 2574  %b20 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 20), align 2575  %b21 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 21), align 2576  %b22 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 22), align 2577  %b23 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 23), align 2578  %b24 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 24), align 2579  %b25 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 25), align 2580  %b26 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 26), align 2581  %b27 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 27), align 2582  %b28 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 28), align 2583  %b29 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 29), align 2584  %b30 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 30), align 2585  %b31 = load i16, ptr getelementptr inbounds ([32 x i16], ptr @b16, i32 0, i64 31), align 2586  %r0  = call i16 @llvm.fshr.i16(i16 %a0 , i16 %a0 , i16 %b0 )587  %r1  = call i16 @llvm.fshr.i16(i16 %a1 , i16 %a1 , i16 %b1 )588  %r2  = call i16 @llvm.fshr.i16(i16 %a2 , i16 %a2 , i16 %b2 )589  %r3  = call i16 @llvm.fshr.i16(i16 %a3 , i16 %a3 , i16 %b3 )590  %r4  = call i16 @llvm.fshr.i16(i16 %a4 , i16 %a4 , i16 %b4 )591  %r5  = call i16 @llvm.fshr.i16(i16 %a5 , i16 %a5 , i16 %b5 )592  %r6  = call i16 @llvm.fshr.i16(i16 %a6 , i16 %a6 , i16 %b6 )593  %r7  = call i16 @llvm.fshr.i16(i16 %a7 , i16 %a7 , i16 %b7 )594  %r8  = call i16 @llvm.fshr.i16(i16 %a8 , i16 %a8 , i16 %b8 )595  %r9  = call i16 @llvm.fshr.i16(i16 %a9 , i16 %a9 , i16 %b9 )596  %r10 = call i16 @llvm.fshr.i16(i16 %a10, i16 %a10, i16 %b10)597  %r11 = call i16 @llvm.fshr.i16(i16 %a11, i16 %a11, i16 %b11)598  %r12 = call i16 @llvm.fshr.i16(i16 %a12, i16 %a12, i16 %b12)599  %r13 = call i16 @llvm.fshr.i16(i16 %a13, i16 %a13, i16 %b13)600  %r14 = call i16 @llvm.fshr.i16(i16 %a14, i16 %a14, i16 %b14)601  %r15 = call i16 @llvm.fshr.i16(i16 %a15, i16 %a15, i16 %b15)602  %r16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %a16, i16 %b16)603  %r17 = call i16 @llvm.fshr.i16(i16 %a17, i16 %a17, i16 %b17)604  %r18 = call i16 @llvm.fshr.i16(i16 %a18, i16 %a18, i16 %b18)605  %r19 = call i16 @llvm.fshr.i16(i16 %a19, i16 %a19, i16 %b19)606  %r20 = call i16 @llvm.fshr.i16(i16 %a20, i16 %a20, i16 %b20)607  %r21 = call i16 @llvm.fshr.i16(i16 %a21, i16 %a21, i16 %b21)608  %r22 = call i16 @llvm.fshr.i16(i16 %a22, i16 %a22, i16 %b22)609  %r23 = call i16 @llvm.fshr.i16(i16 %a23, i16 %a23, i16 %b23)610  %r24 = call i16 @llvm.fshr.i16(i16 %a24, i16 %a24, i16 %b24)611  %r25 = call i16 @llvm.fshr.i16(i16 %a25, i16 %a25, i16 %b25)612  %r26 = call i16 @llvm.fshr.i16(i16 %a26, i16 %a26, i16 %b26)613  %r27 = call i16 @llvm.fshr.i16(i16 %a27, i16 %a27, i16 %b27)614  %r28 = call i16 @llvm.fshr.i16(i16 %a28, i16 %a28, i16 %b28)615  %r29 = call i16 @llvm.fshr.i16(i16 %a29, i16 %a29, i16 %b29)616  %r30 = call i16 @llvm.fshr.i16(i16 %a30, i16 %a30, i16 %b30)617  %r31 = call i16 @llvm.fshr.i16(i16 %a31, i16 %a31, i16 %b31)618  store i16 %r0 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 0 ), align 2619  store i16 %r1 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 1 ), align 2620  store i16 %r2 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 2 ), align 2621  store i16 %r3 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 3 ), align 2622  store i16 %r4 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 4 ), align 2623  store i16 %r5 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 5 ), align 2624  store i16 %r6 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 6 ), align 2625  store i16 %r7 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 7 ), align 2626  store i16 %r8 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 8 ), align 2627  store i16 %r9 , ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 9 ), align 2628  store i16 %r10, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 10), align 2629  store i16 %r11, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 11), align 2630  store i16 %r12, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 12), align 2631  store i16 %r13, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 13), align 2632  store i16 %r14, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 14), align 2633  store i16 %r15, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 15), align 2634  store i16 %r16, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 16), align 2635  store i16 %r17, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 17), align 2636  store i16 %r18, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 18), align 2637  store i16 %r19, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 19), align 2638  store i16 %r20, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 20), align 2639  store i16 %r21, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 21), align 2640  store i16 %r22, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 22), align 2641  store i16 %r23, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 23), align 2642  store i16 %r24, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 24), align 2643  store i16 %r25, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 25), align 2644  store i16 %r26, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 26), align 2645  store i16 %r27, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 27), align 2646  store i16 %r28, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 28), align 2647  store i16 %r29, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 29), align 2648  store i16 %r30, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 30), align 2649  store i16 %r31, ptr getelementptr inbounds ([32 x i16], ptr @d16, i32 0, i64 31), align 2650  ret void651}652 653define void @fshr_v64i8() {654; SSE-LABEL: @fshr_v64i8(655; SSE-NEXT:    [[TMP1:%.*]] = load <16 x i8>, ptr @a8, align 1656; SSE-NEXT:    [[TMP2:%.*]] = load <16 x i8>, ptr @b8, align 1657; SSE-NEXT:    [[TMP3:%.*]] = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> [[TMP1]], <16 x i8> [[TMP1]], <16 x i8> [[TMP2]])658; SSE-NEXT:    store <16 x i8> [[TMP3]], ptr @d8, align 1659; SSE-NEXT:    [[TMP4:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1660; SSE-NEXT:    [[TMP5:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1661; SSE-NEXT:    [[TMP6:%.*]] = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> [[TMP4]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]])662; SSE-NEXT:    store <16 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 16), align 1663; SSE-NEXT:    [[TMP7:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1664; SSE-NEXT:    [[TMP8:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1665; SSE-NEXT:    [[TMP9:%.*]] = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> [[TMP7]], <16 x i8> [[TMP7]], <16 x i8> [[TMP8]])666; SSE-NEXT:    store <16 x i8> [[TMP9]], ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 32), align 1667; SSE-NEXT:    [[TMP10:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1668; SSE-NEXT:    [[TMP11:%.*]] = load <16 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1669; SSE-NEXT:    [[TMP12:%.*]] = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> [[TMP10]], <16 x i8> [[TMP10]], <16 x i8> [[TMP11]])670; SSE-NEXT:    store <16 x i8> [[TMP12]], ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 48), align 1671; SSE-NEXT:    ret void672;673; AVX-LABEL: @fshr_v64i8(674; AVX-NEXT:    [[TMP1:%.*]] = load <32 x i8>, ptr @a8, align 1675; AVX-NEXT:    [[TMP2:%.*]] = load <32 x i8>, ptr @b8, align 1676; AVX-NEXT:    [[TMP3:%.*]] = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> [[TMP1]], <32 x i8> [[TMP1]], <32 x i8> [[TMP2]])677; AVX-NEXT:    store <32 x i8> [[TMP3]], ptr @d8, align 1678; AVX-NEXT:    [[TMP4:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1679; AVX-NEXT:    [[TMP5:%.*]] = load <32 x i8>, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1680; AVX-NEXT:    [[TMP6:%.*]] = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> [[TMP4]], <32 x i8> [[TMP4]], <32 x i8> [[TMP5]])681; AVX-NEXT:    store <32 x i8> [[TMP6]], ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 32), align 1682; AVX-NEXT:    ret void683;684; AVX512-LABEL: @fshr_v64i8(685; AVX512-NEXT:    [[TMP1:%.*]] = load <64 x i8>, ptr @a8, align 1686; AVX512-NEXT:    [[TMP2:%.*]] = load <64 x i8>, ptr @b8, align 1687; AVX512-NEXT:    [[TMP3:%.*]] = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> [[TMP1]], <64 x i8> [[TMP1]], <64 x i8> [[TMP2]])688; AVX512-NEXT:    store <64 x i8> [[TMP3]], ptr @d8, align 1689; AVX512-NEXT:    ret void690;691; AVX512VBMI2-LABEL: @fshr_v64i8(692; AVX512VBMI2-NEXT:    [[TMP1:%.*]] = load <64 x i8>, ptr @a8, align 1693; AVX512VBMI2-NEXT:    [[TMP2:%.*]] = load <64 x i8>, ptr @b8, align 1694; AVX512VBMI2-NEXT:    [[TMP3:%.*]] = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> [[TMP1]], <64 x i8> [[TMP1]], <64 x i8> [[TMP2]])695; AVX512VBMI2-NEXT:    store <64 x i8> [[TMP3]], ptr @d8, align 1696; AVX512VBMI2-NEXT:    ret void697;698  %a0  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 0 ), align 1699  %a1  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 1 ), align 1700  %a2  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 2 ), align 1701  %a3  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 3 ), align 1702  %a4  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 4 ), align 1703  %a5  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 5 ), align 1704  %a6  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 6 ), align 1705  %a7  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 7 ), align 1706  %a8  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 8 ), align 1707  %a9  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 9 ), align 1708  %a10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 10), align 1709  %a11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 11), align 1710  %a12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 12), align 1711  %a13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 13), align 1712  %a14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 14), align 1713  %a15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 15), align 1714  %a16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 16), align 1715  %a17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 17), align 1716  %a18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 18), align 1717  %a19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 19), align 1718  %a20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 20), align 1719  %a21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 21), align 1720  %a22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 22), align 1721  %a23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 23), align 1722  %a24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 24), align 1723  %a25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 25), align 1724  %a26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 26), align 1725  %a27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 27), align 1726  %a28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 28), align 1727  %a29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 29), align 1728  %a30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 30), align 1729  %a31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 31), align 1730  %a32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 32), align 1731  %a33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 33), align 1732  %a34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 34), align 1733  %a35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 35), align 1734  %a36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 36), align 1735  %a37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 37), align 1736  %a38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 38), align 1737  %a39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 39), align 1738  %a40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 40), align 1739  %a41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 41), align 1740  %a42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 42), align 1741  %a43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 43), align 1742  %a44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 44), align 1743  %a45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 45), align 1744  %a46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 46), align 1745  %a47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 47), align 1746  %a48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 48), align 1747  %a49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 49), align 1748  %a50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 50), align 1749  %a51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 51), align 1750  %a52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 52), align 1751  %a53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 53), align 1752  %a54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 54), align 1753  %a55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 55), align 1754  %a56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 56), align 1755  %a57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 57), align 1756  %a58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 58), align 1757  %a59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 59), align 1758  %a60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 60), align 1759  %a61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 61), align 1760  %a62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 62), align 1761  %a63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @a8, i32 0, i64 63), align 1762  %b0  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 0 ), align 1763  %b1  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 1 ), align 1764  %b2  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 2 ), align 1765  %b3  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 3 ), align 1766  %b4  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 4 ), align 1767  %b5  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 5 ), align 1768  %b6  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 6 ), align 1769  %b7  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 7 ), align 1770  %b8  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 8 ), align 1771  %b9  = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 9 ), align 1772  %b10 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 10), align 1773  %b11 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 11), align 1774  %b12 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 12), align 1775  %b13 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 13), align 1776  %b14 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 14), align 1777  %b15 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 15), align 1778  %b16 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 16), align 1779  %b17 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 17), align 1780  %b18 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 18), align 1781  %b19 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 19), align 1782  %b20 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 20), align 1783  %b21 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 21), align 1784  %b22 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 22), align 1785  %b23 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 23), align 1786  %b24 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 24), align 1787  %b25 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 25), align 1788  %b26 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 26), align 1789  %b27 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 27), align 1790  %b28 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 28), align 1791  %b29 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 29), align 1792  %b30 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 30), align 1793  %b31 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 31), align 1794  %b32 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 32), align 1795  %b33 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 33), align 1796  %b34 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 34), align 1797  %b35 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 35), align 1798  %b36 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 36), align 1799  %b37 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 37), align 1800  %b38 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 38), align 1801  %b39 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 39), align 1802  %b40 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 40), align 1803  %b41 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 41), align 1804  %b42 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 42), align 1805  %b43 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 43), align 1806  %b44 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 44), align 1807  %b45 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 45), align 1808  %b46 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 46), align 1809  %b47 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 47), align 1810  %b48 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 48), align 1811  %b49 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 49), align 1812  %b50 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 50), align 1813  %b51 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 51), align 1814  %b52 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 52), align 1815  %b53 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 53), align 1816  %b54 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 54), align 1817  %b55 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 55), align 1818  %b56 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 56), align 1819  %b57 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 57), align 1820  %b58 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 58), align 1821  %b59 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 59), align 1822  %b60 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 60), align 1823  %b61 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 61), align 1824  %b62 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 62), align 1825  %b63 = load i8, ptr getelementptr inbounds ([64 x i8], ptr @b8, i32 0, i64 63), align 1826  %r0  = call i8 @llvm.fshr.i8(i8 %a0 , i8 %a0 , i8 %b0 )827  %r1  = call i8 @llvm.fshr.i8(i8 %a1 , i8 %a1 , i8 %b1 )828  %r2  = call i8 @llvm.fshr.i8(i8 %a2 , i8 %a2 , i8 %b2 )829  %r3  = call i8 @llvm.fshr.i8(i8 %a3 , i8 %a3 , i8 %b3 )830  %r4  = call i8 @llvm.fshr.i8(i8 %a4 , i8 %a4 , i8 %b4 )831  %r5  = call i8 @llvm.fshr.i8(i8 %a5 , i8 %a5 , i8 %b5 )832  %r6  = call i8 @llvm.fshr.i8(i8 %a6 , i8 %a6 , i8 %b6 )833  %r7  = call i8 @llvm.fshr.i8(i8 %a7 , i8 %a7 , i8 %b7 )834  %r8  = call i8 @llvm.fshr.i8(i8 %a8 , i8 %a8 , i8 %b8 )835  %r9  = call i8 @llvm.fshr.i8(i8 %a9 , i8 %a9 , i8 %b9 )836  %r10 = call i8 @llvm.fshr.i8(i8 %a10, i8 %a10, i8 %b10)837  %r11 = call i8 @llvm.fshr.i8(i8 %a11, i8 %a11, i8 %b11)838  %r12 = call i8 @llvm.fshr.i8(i8 %a12, i8 %a12, i8 %b12)839  %r13 = call i8 @llvm.fshr.i8(i8 %a13, i8 %a13, i8 %b13)840  %r14 = call i8 @llvm.fshr.i8(i8 %a14, i8 %a14, i8 %b14)841  %r15 = call i8 @llvm.fshr.i8(i8 %a15, i8 %a15, i8 %b15)842  %r16 = call i8 @llvm.fshr.i8(i8 %a16, i8 %a16, i8 %b16)843  %r17 = call i8 @llvm.fshr.i8(i8 %a17, i8 %a17, i8 %b17)844  %r18 = call i8 @llvm.fshr.i8(i8 %a18, i8 %a18, i8 %b18)845  %r19 = call i8 @llvm.fshr.i8(i8 %a19, i8 %a19, i8 %b19)846  %r20 = call i8 @llvm.fshr.i8(i8 %a20, i8 %a20, i8 %b20)847  %r21 = call i8 @llvm.fshr.i8(i8 %a21, i8 %a21, i8 %b21)848  %r22 = call i8 @llvm.fshr.i8(i8 %a22, i8 %a22, i8 %b22)849  %r23 = call i8 @llvm.fshr.i8(i8 %a23, i8 %a23, i8 %b23)850  %r24 = call i8 @llvm.fshr.i8(i8 %a24, i8 %a24, i8 %b24)851  %r25 = call i8 @llvm.fshr.i8(i8 %a25, i8 %a25, i8 %b25)852  %r26 = call i8 @llvm.fshr.i8(i8 %a26, i8 %a26, i8 %b26)853  %r27 = call i8 @llvm.fshr.i8(i8 %a27, i8 %a27, i8 %b27)854  %r28 = call i8 @llvm.fshr.i8(i8 %a28, i8 %a28, i8 %b28)855  %r29 = call i8 @llvm.fshr.i8(i8 %a29, i8 %a29, i8 %b29)856  %r30 = call i8 @llvm.fshr.i8(i8 %a30, i8 %a30, i8 %b30)857  %r31 = call i8 @llvm.fshr.i8(i8 %a31, i8 %a31, i8 %b31)858  %r32 = call i8 @llvm.fshr.i8(i8 %a32, i8 %a32, i8 %b32)859  %r33 = call i8 @llvm.fshr.i8(i8 %a33, i8 %a33, i8 %b33)860  %r34 = call i8 @llvm.fshr.i8(i8 %a34, i8 %a34, i8 %b34)861  %r35 = call i8 @llvm.fshr.i8(i8 %a35, i8 %a35, i8 %b35)862  %r36 = call i8 @llvm.fshr.i8(i8 %a36, i8 %a36, i8 %b36)863  %r37 = call i8 @llvm.fshr.i8(i8 %a37, i8 %a37, i8 %b37)864  %r38 = call i8 @llvm.fshr.i8(i8 %a38, i8 %a38, i8 %b38)865  %r39 = call i8 @llvm.fshr.i8(i8 %a39, i8 %a39, i8 %b39)866  %r40 = call i8 @llvm.fshr.i8(i8 %a40, i8 %a40, i8 %b40)867  %r41 = call i8 @llvm.fshr.i8(i8 %a41, i8 %a41, i8 %b41)868  %r42 = call i8 @llvm.fshr.i8(i8 %a42, i8 %a42, i8 %b42)869  %r43 = call i8 @llvm.fshr.i8(i8 %a43, i8 %a43, i8 %b43)870  %r44 = call i8 @llvm.fshr.i8(i8 %a44, i8 %a44, i8 %b44)871  %r45 = call i8 @llvm.fshr.i8(i8 %a45, i8 %a45, i8 %b45)872  %r46 = call i8 @llvm.fshr.i8(i8 %a46, i8 %a46, i8 %b46)873  %r47 = call i8 @llvm.fshr.i8(i8 %a47, i8 %a47, i8 %b47)874  %r48 = call i8 @llvm.fshr.i8(i8 %a48, i8 %a48, i8 %b48)875  %r49 = call i8 @llvm.fshr.i8(i8 %a49, i8 %a49, i8 %b49)876  %r50 = call i8 @llvm.fshr.i8(i8 %a50, i8 %a50, i8 %b50)877  %r51 = call i8 @llvm.fshr.i8(i8 %a51, i8 %a51, i8 %b51)878  %r52 = call i8 @llvm.fshr.i8(i8 %a52, i8 %a52, i8 %b52)879  %r53 = call i8 @llvm.fshr.i8(i8 %a53, i8 %a53, i8 %b53)880  %r54 = call i8 @llvm.fshr.i8(i8 %a54, i8 %a54, i8 %b54)881  %r55 = call i8 @llvm.fshr.i8(i8 %a55, i8 %a55, i8 %b55)882  %r56 = call i8 @llvm.fshr.i8(i8 %a56, i8 %a56, i8 %b56)883  %r57 = call i8 @llvm.fshr.i8(i8 %a57, i8 %a57, i8 %b57)884  %r58 = call i8 @llvm.fshr.i8(i8 %a58, i8 %a58, i8 %b58)885  %r59 = call i8 @llvm.fshr.i8(i8 %a59, i8 %a59, i8 %b59)886  %r60 = call i8 @llvm.fshr.i8(i8 %a60, i8 %a60, i8 %b60)887  %r61 = call i8 @llvm.fshr.i8(i8 %a61, i8 %a61, i8 %b61)888  %r62 = call i8 @llvm.fshr.i8(i8 %a62, i8 %a62, i8 %b62)889  %r63 = call i8 @llvm.fshr.i8(i8 %a63, i8 %a63, i8 %b63)890  store i8 %r0 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 0 ), align 1891  store i8 %r1 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 1 ), align 1892  store i8 %r2 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 2 ), align 1893  store i8 %r3 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 3 ), align 1894  store i8 %r4 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 4 ), align 1895  store i8 %r5 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 5 ), align 1896  store i8 %r6 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 6 ), align 1897  store i8 %r7 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 7 ), align 1898  store i8 %r8 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 8 ), align 1899  store i8 %r9 , ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 9 ), align 1900  store i8 %r10, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 10), align 1901  store i8 %r11, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 11), align 1902  store i8 %r12, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 12), align 1903  store i8 %r13, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 13), align 1904  store i8 %r14, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 14), align 1905  store i8 %r15, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 15), align 1906  store i8 %r16, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 16), align 1907  store i8 %r17, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 17), align 1908  store i8 %r18, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 18), align 1909  store i8 %r19, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 19), align 1910  store i8 %r20, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 20), align 1911  store i8 %r21, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 21), align 1912  store i8 %r22, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 22), align 1913  store i8 %r23, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 23), align 1914  store i8 %r24, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 24), align 1915  store i8 %r25, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 25), align 1916  store i8 %r26, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 26), align 1917  store i8 %r27, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 27), align 1918  store i8 %r28, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 28), align 1919  store i8 %r29, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 29), align 1920  store i8 %r30, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 30), align 1921  store i8 %r31, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 31), align 1922  store i8 %r32, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 32), align 1923  store i8 %r33, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 33), align 1924  store i8 %r34, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 34), align 1925  store i8 %r35, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 35), align 1926  store i8 %r36, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 36), align 1927  store i8 %r37, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 37), align 1928  store i8 %r38, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 38), align 1929  store i8 %r39, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 39), align 1930  store i8 %r40, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 40), align 1931  store i8 %r41, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 41), align 1932  store i8 %r42, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 42), align 1933  store i8 %r43, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 43), align 1934  store i8 %r44, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 44), align 1935  store i8 %r45, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 45), align 1936  store i8 %r46, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 46), align 1937  store i8 %r47, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 47), align 1938  store i8 %r48, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 48), align 1939  store i8 %r49, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 49), align 1940  store i8 %r50, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 50), align 1941  store i8 %r51, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 51), align 1942  store i8 %r52, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 52), align 1943  store i8 %r53, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 53), align 1944  store i8 %r54, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 54), align 1945  store i8 %r55, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 55), align 1946  store i8 %r56, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 56), align 1947  store i8 %r57, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 57), align 1948  store i8 %r58, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 58), align 1949  store i8 %r59, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 59), align 1950  store i8 %r60, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 60), align 1951  store i8 %r61, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 61), align 1952  store i8 %r62, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 62), align 1953  store i8 %r63, ptr getelementptr inbounds ([64 x i8], ptr @d8, i32 0, i64 63), align 1954  ret void955}956 957define void @fshr_v2i32() {958; CHECK-LABEL: @fshr_v2i32(959; CHECK-NEXT:    [[A0:%.*]] = load i32, ptr @a32, align 4960; CHECK-NEXT:    [[A1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1), align 4961; CHECK-NEXT:    [[B0:%.*]] = load i32, ptr @b32, align 4962; CHECK-NEXT:    [[B1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1), align 4963; CHECK-NEXT:    [[R0:%.*]] = call i32 @llvm.fshr.i32(i32 [[A0]], i32 [[A0]], i32 [[B0]])964; CHECK-NEXT:    [[R1:%.*]] = call i32 @llvm.fshr.i32(i32 [[A1]], i32 [[A1]], i32 [[B1]])965; CHECK-NEXT:    store i32 [[R0]], ptr @d32, align 4966; CHECK-NEXT:    store i32 [[R1]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 1), align 4967; CHECK-NEXT:    ret void968;969; AVX512VBMI2-LABEL: @fshr_v2i32(970; AVX512VBMI2-NEXT:    [[TMP1:%.*]] = load <2 x i32>, ptr @a32, align 4971; AVX512VBMI2-NEXT:    [[TMP2:%.*]] = load <2 x i32>, ptr @b32, align 4972; AVX512VBMI2-NEXT:    [[TMP3:%.*]] = call <2 x i32> @llvm.fshr.v2i32(<2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> [[TMP2]])973; AVX512VBMI2-NEXT:    store <2 x i32> [[TMP3]], ptr @d32, align 4974; AVX512VBMI2-NEXT:    ret void975;976  %a0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 0 ), align 4977  %a1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1 ), align 4978  %b0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 0 ), align 4979  %b1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @b32, i32 0, i64 1 ), align 4980  %r0  = call i32 @llvm.fshr.i32(i32 %a0 , i32 %a0 , i32 %b0 )981  %r1  = call i32 @llvm.fshr.i32(i32 %a1 , i32 %a1 , i32 %b1 )982  store i32 %r0 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 0 ), align 4983  store i32 %r1 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 1 ), align 4984  ret void985}986 987; PR63980988define void @fshr_v2i32_uniformconst() {989; CHECK-LABEL: @fshr_v2i32_uniformconst(990; CHECK-NEXT:    [[A0:%.*]] = load i32, ptr @a32, align 4991; CHECK-NEXT:    [[A1:%.*]] = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1), align 4992; CHECK-NEXT:    [[R0:%.*]] = call i32 @llvm.fshr.i32(i32 [[A0]], i32 [[A0]], i32 1)993; CHECK-NEXT:    [[R1:%.*]] = call i32 @llvm.fshr.i32(i32 [[A1]], i32 [[A1]], i32 1)994; CHECK-NEXT:    store i32 [[R0]], ptr @d32, align 4995; CHECK-NEXT:    store i32 [[R1]], ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 1), align 4996; CHECK-NEXT:    ret void997;998; AVX512VBMI2-LABEL: @fshr_v2i32_uniformconst(999; AVX512VBMI2-NEXT:    [[TMP1:%.*]] = load <2 x i32>, ptr @a32, align 41000; AVX512VBMI2-NEXT:    [[TMP2:%.*]] = call <2 x i32> @llvm.fshr.v2i32(<2 x i32> [[TMP1]], <2 x i32> [[TMP1]], <2 x i32> splat (i32 1))1001; AVX512VBMI2-NEXT:    store <2 x i32> [[TMP2]], ptr @d32, align 41002; AVX512VBMI2-NEXT:    ret void1003;1004  %a0  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 0 ), align 41005  %a1  = load i32, ptr getelementptr inbounds ([16 x i32], ptr @a32, i32 0, i64 1 ), align 41006  %r0  = call i32 @llvm.fshr.i32(i32 %a0 , i32 %a0 , i32 1 )1007  %r1  = call i32 @llvm.fshr.i32(i32 %a1 , i32 %a1 , i32 1 )1008  store i32 %r0 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 0 ), align 41009  store i32 %r1 , ptr getelementptr inbounds ([16 x i32], ptr @d32, i32 0, i64 1 ), align 41010  ret void1011}1012