| blob | 3e1c043138322f6dc2ef9c344474248bcaca8f0e |
1 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2 ; RUN: opt --slp-vectorizer -mtriple=x86_64-unknown-linux-gnu -o - -S < %s | FileCheck %s
4 @b = common dso_local global i32* null, align 8
5 @e = common dso_local global float 0.000000e+00, align 4
6 @c = common dso_local global float 0.000000e+00, align 4
7 @g = common dso_local global float 0.000000e+00, align 4
8 @d = common dso_local global float 0.000000e+00, align 4
9 @f = common dso_local global float 0.000000e+00, align 4
10 @a = common dso_local global i32 0, align 4
11 @h = common dso_local global float 0.000000e+00, align 4
13 define dso_local void @j() local_unnamed_addr {
14 ; CHECK-LABEL: @j(
15 ; CHECK-NEXT: entry:
16 ; CHECK-NEXT: [[TMP0:%.*]] = load i32*, i32** @b, align 8
17 ; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 4
18 ; CHECK-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 12
19 ; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 5
20 ; CHECK-NEXT: [[TMP1:%.*]] = bitcast i32* [[ARRAYIDX]] to <2 x i32>*
21 ; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i32>, <2 x i32>* [[TMP1]], align 4
22 ; CHECK-NEXT: [[ARRAYIDX3:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 13
23 ; CHECK-NEXT: [[TMP3:%.*]] = bitcast i32* [[ARRAYIDX1]] to <2 x i32>*
24 ; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i32>, <2 x i32>* [[TMP3]], align 4
25 ; CHECK-NEXT: [[TMP5:%.*]] = add nsw <2 x i32> [[TMP4]], [[TMP2]]
26 ; CHECK-NEXT: [[TMP6:%.*]] = sitofp <2 x i32> [[TMP5]] to <2 x float>
27 ; CHECK-NEXT: [[TMP7:%.*]] = fmul <2 x float> [[TMP6]], <float 1.000000e+01, float 1.000000e+01>
28 ; CHECK-NEXT: [[TMP8:%.*]] = fsub <2 x float> <float 1.000000e+00, float 0.000000e+00>, [[TMP7]]
29 ; CHECK-NEXT: [[SHUFFLE:%.*]] = shufflevector <2 x float> [[TMP8]], <2 x float> poison, <4 x i32> <i32 0, i32 0, i32 1, i32 1>
30 ; CHECK-NEXT: [[TMP9:%.*]] = extractelement <4 x float> [[SHUFFLE]], i32 0
31 ; CHECK-NEXT: store float [[TMP9]], float* @g, align 4
32 ; CHECK-NEXT: [[TMP10:%.*]] = fadd <4 x float> [[SHUFFLE]], <float -1.000000e+00, float 1.000000e+00, float -1.000000e+00, float 1.000000e+00>
33 ; CHECK-NEXT: [[TMP11:%.*]] = extractelement <4 x float> [[TMP10]], i32 3
34 ; CHECK-NEXT: store float [[TMP11]], float* @c, align 4
35 ; CHECK-NEXT: [[TMP12:%.*]] = extractelement <4 x float> [[TMP10]], i32 2
36 ; CHECK-NEXT: store float [[TMP12]], float* @d, align 4
37 ; CHECK-NEXT: [[TMP13:%.*]] = extractelement <4 x float> [[TMP10]], i32 1
38 ; CHECK-NEXT: store float [[TMP13]], float* @e, align 4
39 ; CHECK-NEXT: [[TMP14:%.*]] = extractelement <4 x float> [[TMP10]], i32 0
40 ; CHECK-NEXT: store float [[TMP14]], float* @f, align 4
41 ; CHECK-NEXT: [[ARRAYIDX15:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 14
42 ; CHECK-NEXT: [[ARRAYIDX18:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 15
43 ; CHECK-NEXT: [[TMP15:%.*]] = load i32, i32* @a, align 4
44 ; CHECK-NEXT: [[CONV19:%.*]] = sitofp i32 [[TMP15]] to float
45 ; CHECK-NEXT: [[TMP16:%.*]] = insertelement <4 x float> <float -1.000000e+00, float -1.000000e+00, float poison, float poison>, float [[CONV19]], i32 2
46 ; CHECK-NEXT: [[TMP17:%.*]] = extractelement <4 x float> [[SHUFFLE]], i32 2
47 ; CHECK-NEXT: [[TMP18:%.*]] = insertelement <4 x float> [[TMP16]], float [[TMP17]], i32 3
48 ; CHECK-NEXT: [[TMP19:%.*]] = fadd <4 x float> [[TMP10]], [[TMP18]]
49 ; CHECK-NEXT: [[TMP20:%.*]] = fsub <4 x float> [[TMP10]], [[TMP18]]
50 ; CHECK-NEXT: [[TMP21:%.*]] = shufflevector <4 x float> [[TMP19]], <4 x float> [[TMP20]], <4 x i32> <i32 0, i32 1, i32 6, i32 7>
51 ; CHECK-NEXT: [[TMP22:%.*]] = fptosi <4 x float> [[TMP21]] to <4 x i32>
52 ; CHECK-NEXT: [[SHUFFLE1:%.*]] = shufflevector <4 x i32> [[TMP22]], <4 x i32> poison, <4 x i32> <i32 2, i32 0, i32 3, i32 1>
53 ; CHECK-NEXT: [[TMP23:%.*]] = bitcast i32* [[ARRAYIDX1]] to <4 x i32>*
54 ; CHECK-NEXT: store <4 x i32> [[SHUFFLE1]], <4 x i32>* [[TMP23]], align 4
55 ; CHECK-NEXT: ret void
56 ;
57 entry:
58 %0 = load i32*, i32** @b, align 8
59 %arrayidx = getelementptr inbounds i32, i32* %0, i64 4
60 %1 = load i32, i32* %arrayidx, align 4
61 %arrayidx1 = getelementptr inbounds i32, i32* %0, i64 12
62 %2 = load i32, i32* %arrayidx1, align 4
63 %add = add nsw i32 %2, %1
64 %conv = sitofp i32 %add to float
65 %mul = fmul float %conv, 1.000000e+01
66 %arrayidx2 = getelementptr inbounds i32, i32* %0, i64 5
67 %3 = load i32, i32* %arrayidx2, align 4
68 %arrayidx3 = getelementptr inbounds i32, i32* %0, i64 13
69 %4 = load i32, i32* %arrayidx3, align 4
70 %add4 = add nsw i32 %4, %3
71 %conv5 = sitofp i32 %add4 to float
72 %mul6 = fmul float %conv5, 1.000000e+01
73 %sub = fsub float 0.000000e+00, %mul6
74 %sub7 = fsub float 1.000000e+00, %mul
75 store float %sub7, float* @g, align 4
76 %add9 = fadd float %sub, 1.000000e+00
77 store float %add9, float* @c, align 4
78 %sub10 = fadd float %sub, -1.000000e+00
79 store float %sub10, float* @d, align 4
80 %add11 = fadd float %sub7, 1.000000e+00
81 store float %add11, float* @e, align 4
82 %sub12 = fadd float %sub7, -1.000000e+00
83 store float %sub12, float* @f, align 4
84 %sub13 = fsub float %add9, %sub
85 %conv14 = fptosi float %sub13 to i32
86 %arrayidx15 = getelementptr inbounds i32, i32* %0, i64 14
87 store i32 %conv14, i32* %arrayidx15, align 4
88 %sub16 = fadd float %add11, -1.000000e+00
89 %conv17 = fptosi float %sub16 to i32
90 %arrayidx18 = getelementptr inbounds i32, i32* %0, i64 15
91 store i32 %conv17, i32* %arrayidx18, align 4
92 %5 = load i32, i32* @a, align 4
93 %conv19 = sitofp i32 %5 to float
94 %sub20 = fsub float %sub10, %conv19
95 %conv21 = fptosi float %sub20 to i32
96 store i32 %conv21, i32* %arrayidx1, align 4
97 %sub23 = fadd float %sub12, -1.000000e+00
98 %conv24 = fptosi float %sub23 to i32
99 store i32 %conv24, i32* %arrayidx3, align 4
100 ret void
101 }