1 ; RUN: llc < %s -mcpu=g5 | FileCheck %s
2 ; RUN: llc < %s -mcpu=g5 | FileCheck %s -check-prefix=CHECK-PC
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-f128:128:128-v128:128:128-n32:64"
4 target triple = "powerpc64-unknown-linux-gnu"
6 define void @foo(float* noalias nocapture %a, float* noalias nocapture %b) #0 {
10 vector.body: ; preds = %vector.body, %vector.ph
11 %index = phi i64 [ 0, %vector.ph ], [ %index.next, %vector.body ]
12 %0 = getelementptr inbounds float* %b, i64 %index
13 %1 = bitcast float* %0 to <4 x float>*
14 %wide.load = load <4 x float>* %1, align 4
15 %.sum11 = or i64 %index, 4
16 %2 = getelementptr float* %b, i64 %.sum11
17 %3 = bitcast float* %2 to <4 x float>*
18 %wide.load8 = load <4 x float>* %3, align 4
19 %4 = fadd <4 x float> %wide.load, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
20 %5 = fadd <4 x float> %wide.load8, <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
21 %6 = getelementptr inbounds float* %a, i64 %index
22 %7 = bitcast float* %6 to <4 x float>*
23 store <4 x float> %4, <4 x float>* %7, align 4
24 %.sum12 = or i64 %index, 4
25 %8 = getelementptr float* %a, i64 %.sum12
26 %9 = bitcast float* %8 to <4 x float>*
27 store <4 x float> %5, <4 x float>* %9, align 4
28 %index.next = add i64 %index, 8
29 %10 = icmp eq i64 %index.next, 16000
30 br i1 %10, label %for.end, label %vector.body
33 ; CHECK: lvx [[CNST:[0-9]+]],
34 ; CHECK-DAG: lvsl [[PC:[0-9]+]], [[B1:[0-9]+]], [[B2:[0-9]+]]
35 ; CHECK-DAG: lvx [[LD1:[0-9]+]], [[B1]], [[B2]]
36 ; CHECK-DAG: add [[B3:[0-9]+]], [[B1]], [[B2]]
37 ; CHECK-DAG: lvx [[LD2:[0-9]+]], [[B3]],
38 ; CHECK-DAG: vperm [[R1:[0-9]+]], [[LD1]], [[LD2]], [[PC]]
39 ; CHECK: vaddfp {{[0-9]+}}, [[R1]], [[CNST]]
47 for.end: ; preds = %vector.body
51 attributes #0 = { nounwind }