1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py 2; RUN: opt < %s -slp-vectorizer -S -mtriple=x86_64-unknown-linux -mcpu=core-avx2 | FileCheck %s 3target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128" 4 5%struct.S = type { [3 x float], [3 x float], [4 x float] } 6 7define i32 @foo(i32 %0, i32* %1, float* %2) { 8; CHECK-LABEL: @foo( 9; CHECK-NEXT: [[T4:%.*]] = alloca [[STRUCT_S:%.*]], align 8 10; CHECK-NEXT: [[T8:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[T4]], i64 0, i32 1 11; CHECK-NEXT: [[T9:%.*]] = getelementptr inbounds [3 x float], [3 x float]* [[T8]], i64 1, i64 0 12; CHECK-NEXT: [[T14:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[T4]], i64 0, i32 1, i64 0 13; CHECK-NEXT: [[T11:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[T4]], i64 0, i32 1, i64 1 14; CHECK-NEXT: [[TMP4:%.*]] = bitcast float* [[T14]] to <2 x float>* 15; CHECK-NEXT: [[TMP5:%.*]] = load <2 x float>, <2 x float>* [[TMP4]], align 4 16; CHECK-NEXT: br label [[T37:%.*]] 17; CHECK: t37: 18; CHECK-NEXT: [[TMP6:%.*]] = phi <2 x float> [ [[TMP5]], [[TMP3:%.*]] ], [ [[T89:%.*]], [[T37]] ] 19; CHECK-NEXT: [[TMP7:%.*]] = fdiv fast <2 x float> <float 1.000000e+00, float 1.000000e+00>, [[TMP6]] 20; CHECK-NEXT: [[SHUFFLE:%.*]] = shufflevector <2 x float> [[TMP7]], <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 1, i32 1> 21; CHECK-NEXT: [[T21:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[T4]], i64 0, i32 2, i64 0 22; CHECK-NEXT: [[T25:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[T4]], i64 0, i32 2, i64 1 23; CHECK-NEXT: [[T31:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[T4]], i64 0, i32 2, i64 2 24; CHECK-NEXT: [[T33:%.*]] = getelementptr inbounds [[STRUCT_S]], %struct.S* [[T4]], i64 0, i32 2, i64 3 25; CHECK-NEXT: [[TMP8:%.*]] = bitcast float* [[T21]] to <4 x float>* 26; CHECK-NEXT: store <4 x float> [[SHUFFLE]], <4 x float>* [[TMP8]], align 4 27; CHECK-NEXT: [[T88:%.*]] = bitcast float* [[T9]] to <2 x float>* 28; CHECK-NEXT: [[T89]] = load <2 x float>, <2 x float>* [[T88]], align 4 29; CHECK-NEXT: br i1 undef, label [[T37]], label [[T55:%.*]] 30; CHECK: t55: 31; CHECK-NEXT: ret i32 0 32; 33 %t4 = alloca %struct.S, align 8 34 %t8 = getelementptr inbounds %struct.S, %struct.S* %t4, i64 0, i32 1 35 %t9 = getelementptr inbounds [3 x float], [3 x float]* %t8, i64 1, i64 0 36 %t14 = getelementptr inbounds %struct.S, %struct.S* %t4, i64 0, i32 1, i64 0 37 %t11 = getelementptr inbounds %struct.S, %struct.S* %t4, i64 0, i32 1, i64 1 38 %t15 = load float, float* %t14, align 4 39 %t16 = load float, float* %t11, align 4 40 br label %t37 41 42t37: 43 44 %t18 = phi float [ %t16, %3 ], [ %x24, %t37 ] 45 %t19 = phi float [ %t15, %3 ], [ %x23, %t37 ] 46 %t20 = fdiv fast float 1.000000e+00, %t19 47 %t24 = fdiv fast float 1.000000e+00, %t18 48 %t21 = getelementptr inbounds %struct.S, %struct.S* %t4, i64 0, i32 2, i64 0 49 %t25 = getelementptr inbounds %struct.S, %struct.S* %t4, i64 0, i32 2, i64 1 50 %t31 = getelementptr inbounds %struct.S, %struct.S* %t4, i64 0, i32 2, i64 2 51 %t33 = getelementptr inbounds %struct.S, %struct.S* %t4, i64 0, i32 2, i64 3 52 store float %t20, float* %t21, align 4 53 store float %t24, float* %t25, align 4 54 store float %t24, float* %t31, align 4 55 store float %t24, float* %t33, align 4 56 %t88 = bitcast float* %t9 to <2 x float>* 57 %t89 = load <2 x float>, <2 x float>* %t88, align 4 58 %x23 = extractelement <2 x float> %t89, i32 0 59 %x24 = extractelement <2 x float> %t89, i32 1 60 br i1 undef, label %t37, label %t55 61 62t55: 63 64 ret i32 0 65 66} 67 68