1 ; RUN: opt < %s -basicaa -slp-vectorizer -dce -S -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s 2 3 target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" 4 target triple = "x86_64-apple-macosx10.8.0" 5 ;CHECK: fextr 6 ;CHECK-NOT: insertelement 7 ;CHECK-NOT: extractelement 8 ;CHECK: fadd <2 x double> 9 ;CHECK: ret void 10 define void @fextr(double* %ptr) { 11 entry: 12 %LD = load <2 x double>, <2 x double>* undef 13 %V0 = extractelement <2 x double> %LD, i32 0 14 %V1 = extractelement <2 x double> %LD, i32 1 15 %P0 = getelementptr inbounds double, double* %ptr, i64 0 16 %P1 = getelementptr inbounds double, double* %ptr, i64 1 17 %A0 = fadd double %V0, 0.0 18 %A1 = fadd double %V1, 1.1 19 store double %A0, double* %P0, align 4 20 store double %A1, double* %P1, align 4 21 ret void 22 } 23 24 ;CHECK: fextr1 25 ;CHECK: insertelement 26 ;CHECK: insertelement 27 ;CHECK: ret void 28 define void @fextr1(double* %ptr) { 29 entry: 30 %LD = load <2 x double>, <2 x double>* undef 31 %V0 = extractelement <2 x double> %LD, i32 0 32 %V1 = extractelement <2 x double> %LD, i32 1 33 %P0 = getelementptr inbounds double, double* %ptr, i64 1 ; <--- incorrect order 34 %P1 = getelementptr inbounds double, double* %ptr, i64 0 35 %A0 = fadd double %V0, 1.2 36 %A1 = fadd double %V1, 3.4 37 store double %A0, double* %P0, align 4 38 store double %A1, double* %P1, align 4 39 ret void 40 } 41 42 ;CHECK: fextr2 43 ;CHECK: insertelement 44 ;CHECK: insertelement 45 ;CHECK: ret void 46 define void @fextr2(double* %ptr) { 47 entry: 48 %LD = load <4 x double>, <4 x double>* undef 49 %V0 = extractelement <4 x double> %LD, i32 0 ; <--- invalid size. 50 %V1 = extractelement <4 x double> %LD, i32 1 51 %P0 = getelementptr inbounds double, double* %ptr, i64 0 52 %P1 = getelementptr inbounds double, double* %ptr, i64 1 53 %A0 = fadd double %V0, 5.5 54 %A1 = fadd double %V1, 6.6 55 store double %A0, double* %P0, align 4 56 store double %A1, double* %P1, align 4 57 ret void 58 } 59 60
