1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py 2; RUN: opt -slp-vectorizer -S -mtriple=x86_64-unknown-linux-gnu < %s | FileCheck %s 3 4; See https://reviews.llvm.org/D83779 5 6define <2 x float> @foo({{float, float}}* %A) { 7; CHECK-LABEL: @foo( 8; CHECK-NEXT: entry: 9; CHECK-NEXT: [[TMP0:%.*]] = bitcast { { float, float } }* [[A:%.*]] to <2 x float>* 10; CHECK-NEXT: [[TMP1:%.*]] = load <2 x float>, <2 x float>* [[TMP0]], align 8 11; CHECK-NEXT: [[TMP2:%.*]] = fmul <2 x float> [[TMP1]], <float 2.000000e+00, float 2.000000e+00> 12; CHECK-NEXT: [[TMP3:%.*]] = extractelement <2 x float> [[TMP2]], i32 1 13; CHECK-NEXT: [[INS1:%.*]] = insertelement <2 x float> undef, float [[TMP3]], i32 1 14; CHECK-NEXT: [[TMP4:%.*]] = extractelement <2 x float> [[TMP2]], i32 0 15; CHECK-NEXT: [[INS0:%.*]] = insertelement <2 x float> [[INS1]], float [[TMP4]], i32 0 16; CHECK-NEXT: ret <2 x float> [[INS0]] 17; 18entry: 19 %0 = bitcast {{float, float}}* %A to <2 x float>* 20 %1 = load <2 x float>, <2 x float>* %0 21 %L0 = extractelement <2 x float> %1, i32 0 22 %L1 = extractelement <2 x float> %1, i32 1 23 %Mul0 = fmul float %L0, 2.000000e+00 24 %Mul1 = fmul float %L1, 2.000000e+00 25 %Ins1 = insertelement <2 x float> undef, float %Mul1, i32 1 26 %Ins0 = insertelement <2 x float> %Ins1, float %Mul0, i32 0 27 ret <2 x float> %Ins0 28} 29 30 31%Struct1Ty = type { i16, i16 } 32%Struct2Ty = type { %Struct1Ty, %Struct1Ty} 33 34define {%Struct2Ty, %Struct2Ty} @StructOfStructOfStruct(i16 *%Ptr) { 35; CHECK-LABEL: @StructOfStructOfStruct( 36; CHECK-NEXT: [[GEP0:%.*]] = getelementptr inbounds i16, i16* [[PTR:%.*]], i64 0 37; CHECK-NEXT: [[GEP1:%.*]] = getelementptr inbounds i16, i16* [[PTR]], i64 1 38; CHECK-NEXT: [[GEP2:%.*]] = getelementptr inbounds i16, i16* [[PTR]], i64 2 39; CHECK-NEXT: [[GEP3:%.*]] = getelementptr inbounds i16, i16* [[PTR]], i64 3 40; CHECK-NEXT: [[GEP4:%.*]] = getelementptr inbounds i16, i16* [[PTR]], i64 4 41; CHECK-NEXT: [[GEP5:%.*]] = getelementptr inbounds i16, i16* [[PTR]], i64 5 42; CHECK-NEXT: [[GEP6:%.*]] = getelementptr inbounds i16, i16* [[PTR]], i64 6 43; CHECK-NEXT: [[GEP7:%.*]] = getelementptr inbounds i16, i16* [[PTR]], i64 7 44; CHECK-NEXT: [[TMP1:%.*]] = bitcast i16* [[GEP0]] to <8 x i16>* 45; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i16>, <8 x i16>* [[TMP1]], align 2 46; CHECK-NEXT: [[TMP3:%.*]] = add <8 x i16> [[TMP2]], <i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 8> 47; CHECK-NEXT: [[TMP4:%.*]] = extractelement <8 x i16> [[TMP3]], i32 1 48; CHECK-NEXT: [[STRUCTIN0:%.*]] = insertvalue [[STRUCT1TY:%.*]] undef, i16 [[TMP4]], 1 49; CHECK-NEXT: [[TMP5:%.*]] = extractelement <8 x i16> [[TMP3]], i32 0 50; CHECK-NEXT: [[STRUCTIN1:%.*]] = insertvalue [[STRUCT1TY]] %StructIn0, i16 [[TMP5]], 0 51; CHECK-NEXT: [[TMP6:%.*]] = extractelement <8 x i16> [[TMP3]], i32 2 52; CHECK-NEXT: [[STRUCTIN2:%.*]] = insertvalue [[STRUCT1TY]] undef, i16 [[TMP6]], 0 53; CHECK-NEXT: [[TMP7:%.*]] = extractelement <8 x i16> [[TMP3]], i32 3 54; CHECK-NEXT: [[STRUCTIN3:%.*]] = insertvalue [[STRUCT1TY]] %StructIn2, i16 [[TMP7]], 1 55; CHECK-NEXT: [[TMP8:%.*]] = extractelement <8 x i16> [[TMP3]], i32 4 56; CHECK-NEXT: [[STRUCTIN4:%.*]] = insertvalue [[STRUCT1TY]] undef, i16 [[TMP8]], 0 57; CHECK-NEXT: [[TMP9:%.*]] = extractelement <8 x i16> [[TMP3]], i32 5 58; CHECK-NEXT: [[STRUCTIN5:%.*]] = insertvalue [[STRUCT1TY]] %StructIn4, i16 [[TMP9]], 1 59; CHECK-NEXT: [[TMP10:%.*]] = extractelement <8 x i16> [[TMP3]], i32 7 60; CHECK-NEXT: [[STRUCTIN6:%.*]] = insertvalue [[STRUCT1TY]] undef, i16 [[TMP10]], 1 61; CHECK-NEXT: [[TMP11:%.*]] = extractelement <8 x i16> [[TMP3]], i32 6 62; CHECK-NEXT: [[STRUCTIN7:%.*]] = insertvalue [[STRUCT1TY]] %StructIn6, i16 [[TMP11]], 0 63; CHECK-NEXT: [[STRUCT2IN0:%.*]] = insertvalue [[STRUCT2TY:%.*]] undef, [[STRUCT1TY]] %StructIn1, 0 64; CHECK-NEXT: [[STRUCT2IN1:%.*]] = insertvalue [[STRUCT2TY]] %Struct2In0, [[STRUCT1TY]] %StructIn3, 1 65; CHECK-NEXT: [[STRUCT2IN2:%.*]] = insertvalue [[STRUCT2TY]] undef, [[STRUCT1TY]] %StructIn5, 0 66; CHECK-NEXT: [[STRUCT2IN3:%.*]] = insertvalue [[STRUCT2TY]] %Struct2In2, [[STRUCT1TY]] %StructIn7, 1 67; CHECK-NEXT: [[RET0:%.*]] = insertvalue { [[STRUCT2TY]], [[STRUCT2TY]] } undef, [[STRUCT2TY]] %Struct2In3, 1 68; CHECK-NEXT: [[RET1:%.*]] = insertvalue { [[STRUCT2TY]], [[STRUCT2TY]] } [[RET0]], [[STRUCT2TY]] %Struct2In1, 0 69; CHECK-NEXT: ret { [[STRUCT2TY]], [[STRUCT2TY]] } [[RET1]] 70; 71 %GEP0 = getelementptr inbounds i16, i16* %Ptr, i64 0 72 %L0 = load i16, i16 * %GEP0 73 %GEP1 = getelementptr inbounds i16, i16* %Ptr, i64 1 74 %L1 = load i16, i16 * %GEP1 75 %GEP2 = getelementptr inbounds i16, i16* %Ptr, i64 2 76 %L2 = load i16, i16 * %GEP2 77 %GEP3 = getelementptr inbounds i16, i16* %Ptr, i64 3 78 %L3 = load i16, i16 * %GEP3 79 %GEP4 = getelementptr inbounds i16, i16* %Ptr, i64 4 80 %L4 = load i16, i16 * %GEP4 81 %GEP5 = getelementptr inbounds i16, i16* %Ptr, i64 5 82 %L5 = load i16, i16 * %GEP5 83 %GEP6 = getelementptr inbounds i16, i16* %Ptr, i64 6 84 %L6 = load i16, i16 * %GEP6 85 %GEP7 = getelementptr inbounds i16, i16* %Ptr, i64 7 86 %L7 = load i16, i16 * %GEP7 87 88 %Fadd0 = add i16 %L0, 1 89 %Fadd1 = add i16 %L1, 2 90 %Fadd2 = add i16 %L2, 3 91 %Fadd3 = add i16 %L3, 4 92 %Fadd4 = add i16 %L4, 5 93 %Fadd5 = add i16 %L5, 6 94 %Fadd6 = add i16 %L6, 7 95 %Fadd7 = add i16 %L7, 8 96 97 %StructIn0 = insertvalue %Struct1Ty undef, i16 %Fadd1, 1 98 %StructIn1 = insertvalue %Struct1Ty %StructIn0, i16 %Fadd0, 0 99 100 %StructIn2 = insertvalue %Struct1Ty undef, i16 %Fadd2, 0 101 %StructIn3 = insertvalue %Struct1Ty %StructIn2, i16 %Fadd3, 1 102 103 %StructIn4 = insertvalue %Struct1Ty undef, i16 %Fadd4, 0 104 %StructIn5 = insertvalue %Struct1Ty %StructIn4, i16 %Fadd5, 1 105 106 %StructIn6 = insertvalue %Struct1Ty undef, i16 %Fadd7, 1 107 %StructIn7 = insertvalue %Struct1Ty %StructIn6, i16 %Fadd6, 0 108 109 %Struct2In0 = insertvalue %Struct2Ty undef, %Struct1Ty %StructIn1, 0 110 %Struct2In1 = insertvalue %Struct2Ty %Struct2In0, %Struct1Ty %StructIn3, 1 111 112 %Struct2In2 = insertvalue %Struct2Ty undef, %Struct1Ty %StructIn5, 0 113 %Struct2In3 = insertvalue %Struct2Ty %Struct2In2, %Struct1Ty %StructIn7, 1 114 115 %Ret0 = insertvalue {%Struct2Ty, %Struct2Ty} undef, %Struct2Ty %Struct2In3, 1 116 %Ret1 = insertvalue {%Struct2Ty, %Struct2Ty} %Ret0, %Struct2Ty %Struct2In1, 0 117 ret {%Struct2Ty, %Struct2Ty} %Ret1 118} 119