1; RUN: llc < %s -mtriple=s390x-linux-gnu -mcpu=z13 | FileCheck %s
2
3; Test that a replicate of a load gets folded to vlrep also in cases where
4; the load has multiple users.
5
6; CHECK-NOT: vrep
7
8
9define double @fun(double* %Vsrc, <2 x double> %T) {
10entry:
11  %Vgep1 = getelementptr double, double* %Vsrc, i64 0
12  %Vld1 = load double, double* %Vgep1
13  %Vgep2 = getelementptr double, double* %Vsrc, i64 1
14  %Vld2 = load double, double* %Vgep2
15  %Vgep3 = getelementptr double, double* %Vsrc, i64 2
16  %Vld3 = load double, double* %Vgep3
17  %Vgep4 = getelementptr double, double* %Vsrc, i64 3
18  %Vld4 = load double, double* %Vgep4
19  %Vgep5 = getelementptr double, double* %Vsrc, i64 4
20  %Vld5 = load double, double* %Vgep5
21  %Vgep6 = getelementptr double, double* %Vsrc, i64 5
22  %Vld6 = load double, double* %Vgep6
23
24  %V19 = insertelement <2 x double> undef, double %Vld1, i32 0
25  %V20 = shufflevector <2 x double> %V19, <2 x double> undef, <2 x i32> zeroinitializer
26  %V21 = insertelement <2 x double> undef, double %Vld4, i32 0
27  %V22 = insertelement <2 x double> %V21, double %Vld5, i32 1
28  %V23 = fmul <2 x double> %V20, %V22
29  %V24 = fadd <2 x double> %T, %V23
30  %V25 = insertelement <2 x double> %V19, double %Vld2, i32 1
31  %V26 = insertelement <2 x double> undef, double %Vld6, i32 0
32  %V27 = insertelement <2 x double> %V26, double %Vld6, i32 1
33  %V28 = fmul <2 x double> %V25, %V27
34  %V29 = fadd <2 x double> %T, %V28
35  %V30 = insertelement <2 x double> undef, double %Vld2, i32 0
36  %V31 = shufflevector <2 x double> %V30, <2 x double> undef, <2 x i32> zeroinitializer
37  %V32 = insertelement <2 x double> undef, double %Vld5, i32 0
38  %V33 = insertelement <2 x double> %V32, double %Vld6, i32 1
39  %V34 = fmul <2 x double> %V31, %V33
40  %V35 = fadd <2 x double> %T, %V34
41  %V36 = insertelement <2 x double> undef, double %Vld3, i32 0
42  %V37 = shufflevector <2 x double> %V36, <2 x double> undef, <2 x i32> zeroinitializer
43  %V38 = fmul <2 x double> %V37, %V22
44  %V39 = fadd <2 x double> %T, %V38
45  %Vmul37 = fmul double %Vld3, %Vld6
46  %Vadd38 = fadd double %Vmul37, %Vmul37
47
48  %VA0 = fadd <2 x double> %V24, %V29
49  %VA1 = fadd <2 x double> %VA0, %V35
50  %VA2 = fadd <2 x double> %VA1, %V39
51
52  %VE0 = extractelement <2 x double> %VA2, i32 0
53  %VS1 = fadd double %VE0, %Vadd38
54
55  ret double %VS1
56}
57