1 ; RUN: opt < %s -basicaa -slp-vectorizer -S -mcpu=corei7-avx | FileCheck %s
2 ; RUN: opt < %s -basicaa -slp-vectorizer -slp-max-reg-size=128 -S -mcpu=corei7-avx | FileCheck %s --check-prefix=V128
4 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
5 target triple = "x86_64-unknown-linux-gnu"
8 ; CHECK: load <4 x double>
9 ; CHECK: fadd <4 x double>
10 ; CHECK: fadd <4 x double>
11 ; CHECK: store <4 x double>
14 ; V128: load <2 x double>
15 ; V128: fadd <2 x double>
16 ; V128: fadd <2 x double>
17 ; V128: store <2 x double>
18 ; V128: load <2 x double>
19 ; V128: fadd <2 x double>
20 ; V128: fadd <2 x double>
21 ; V128: store <2 x double>
23 define void @foo(double* %x) {
24 %1 = load double, double* %x, align 8
25 %2 = fadd double %1, %1
26 %3 = fadd double %2, %1
27 store double %3, double* %x, align 8
28 %4 = getelementptr inbounds double, double* %x, i64 1
29 %5 = load double, double* %4, align 8
30 %6 = fadd double %5, %5
31 %7 = fadd double %6, %5
32 store double %7, double* %4, align 8
33 %8 = getelementptr inbounds double, double* %x, i64 2
34 %9 = load double, double* %8, align 8
35 %10 = fadd double %9, %9
36 %11 = fadd double %10, %9
37 store double %11, double* %8, align 8
38 %12 = getelementptr inbounds double, double* %x, i64 3
39 %13 = load double, double* %12, align 8
40 %14 = fadd double %13, %13
41 %15 = fadd double %14, %13
42 store double %15, double* %12, align 8