• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -slp-vectorizer -slp-vectorizer -mattr=+sse2 -S | FileCheck %s --check-prefix=SSE
3; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -slp-vectorizer -slp-vectorizer -mattr=+avx  -S | FileCheck %s --check-prefix=AVX
4; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -slp-vectorizer -slp-vectorizer -mattr=+avx2 -S | FileCheck %s --check-prefix=AVX
5
6%class.1 = type { %class.2 }
7%class.2 = type { %"class.3" }
8%"class.3" = type { %"struct.1", i64 }
9%"struct.1" = type { [8 x i64] }
10
11$_ZN1C10SwitchModeEv = comdat any
12
13; Function Attrs: uwtable
14define void @_ZN1C10SwitchModeEv() local_unnamed_addr #0 comdat align 2 {
15; SSE-LABEL: @_ZN1C10SwitchModeEv(
16; SSE-NEXT:  for.body.lr.ph.i:
17; SSE-NEXT:    [[OR_1:%.*]] = or i64 undef, 1
18; SSE-NEXT:    store i64 [[OR_1]], i64* undef, align 8
19; SSE-NEXT:    [[FOO_1:%.*]] = getelementptr inbounds [[CLASS_1:%.*]], %class.1* undef, i64 0, i32 0, i32 0, i32 0, i32 0, i64 0
20; SSE-NEXT:    [[FOO_3:%.*]] = load i64, i64* [[FOO_1]], align 8
21; SSE-NEXT:    [[FOO_2:%.*]] = getelementptr inbounds [[CLASS_1]], %class.1* undef, i64 0, i32 0, i32 0, i32 0, i32 0, i64 1
22; SSE-NEXT:    [[FOO_4:%.*]] = load i64, i64* [[FOO_2]], align 8
23; SSE-NEXT:    [[BAR5:%.*]] = load i64, i64* undef, align 8
24; SSE-NEXT:    [[AND_2:%.*]] = and i64 [[OR_1]], [[FOO_3]]
25; SSE-NEXT:    [[AND_1:%.*]] = and i64 [[BAR5]], [[FOO_4]]
26; SSE-NEXT:    [[BAR3:%.*]] = getelementptr inbounds [[CLASS_2:%.*]], %class.2* undef, i64 0, i32 0, i32 0, i32 0, i64 0
27; SSE-NEXT:    store i64 [[AND_2]], i64* [[BAR3]], align 8
28; SSE-NEXT:    [[BAR4:%.*]] = getelementptr inbounds [[CLASS_2]], %class.2* undef, i64 0, i32 0, i32 0, i32 0, i64 1
29; SSE-NEXT:    store i64 [[AND_1]], i64* [[BAR4]], align 8
30; SSE-NEXT:    ret void
31;
32; AVX-LABEL: @_ZN1C10SwitchModeEv(
33; AVX-NEXT:  for.body.lr.ph.i:
34; AVX-NEXT:    [[OR_1:%.*]] = or i64 undef, 1
35; AVX-NEXT:    store i64 [[OR_1]], i64* undef, align 8
36; AVX-NEXT:    [[FOO_1:%.*]] = getelementptr inbounds [[CLASS_1:%.*]], %class.1* undef, i64 0, i32 0, i32 0, i32 0, i32 0, i64 0
37; AVX-NEXT:    [[FOO_2:%.*]] = getelementptr inbounds [[CLASS_1]], %class.1* undef, i64 0, i32 0, i32 0, i32 0, i32 0, i64 1
38; AVX-NEXT:    [[TMP0:%.*]] = bitcast i64* [[FOO_1]] to <2 x i64>*
39; AVX-NEXT:    [[TMP1:%.*]] = load <2 x i64>, <2 x i64>* [[TMP0]], align 8
40; AVX-NEXT:    [[BAR5:%.*]] = load i64, i64* undef, align 8
41; AVX-NEXT:    [[TMP2:%.*]] = insertelement <2 x i64> undef, i64 [[OR_1]], i32 0
42; AVX-NEXT:    [[TMP3:%.*]] = insertelement <2 x i64> [[TMP2]], i64 [[BAR5]], i32 1
43; AVX-NEXT:    [[TMP4:%.*]] = and <2 x i64> [[TMP3]], [[TMP1]]
44; AVX-NEXT:    [[BAR3:%.*]] = getelementptr inbounds [[CLASS_2:%.*]], %class.2* undef, i64 0, i32 0, i32 0, i32 0, i64 0
45; AVX-NEXT:    [[BAR4:%.*]] = getelementptr inbounds [[CLASS_2]], %class.2* undef, i64 0, i32 0, i32 0, i32 0, i64 1
46; AVX-NEXT:    [[TMP5:%.*]] = bitcast i64* [[BAR3]] to <2 x i64>*
47; AVX-NEXT:    store <2 x i64> [[TMP4]], <2 x i64>* [[TMP5]], align 8
48; AVX-NEXT:    ret void
49;
50for.body.lr.ph.i:
51  %or.1 = or i64 undef, 1
52  store i64 %or.1, i64* undef, align 8
53  %foo.1 = getelementptr inbounds %class.1, %class.1* undef, i64 0, i32 0, i32 0, i32 0, i32 0, i64 0
54  %foo.3 = load i64, i64* %foo.1, align 8
55  %foo.2 = getelementptr inbounds %class.1, %class.1* undef, i64 0, i32 0, i32 0, i32 0, i32 0, i64 1
56  %foo.4 = load i64, i64* %foo.2, align 8
57  %bar5 = load i64, i64* undef, align 8
58  %and.2 = and i64 %or.1, %foo.3
59  %and.1 = and i64 %bar5, %foo.4
60  %bar3 = getelementptr inbounds %class.2, %class.2* undef, i64 0, i32 0, i32 0, i32 0, i64 0
61  store i64 %and.2, i64* %bar3, align 8
62  %bar4 = getelementptr inbounds %class.2, %class.2* undef, i64 0, i32 0, i32 0, i32 0, i64 1
63  store i64 %and.1, i64* %bar4, align 8
64  ret void
65}
66
67; Function Attrs: norecurse nounwind uwtable
68define void @pr35497() local_unnamed_addr #0 {
69; SSE-LABEL: @pr35497(
70; SSE-NEXT:  entry:
71; SSE-NEXT:    [[TMP0:%.*]] = load i64, i64* undef, align 1
72; SSE-NEXT:    [[AND:%.*]] = shl i64 [[TMP0]], 2
73; SSE-NEXT:    [[SHL:%.*]] = and i64 [[AND]], 20
74; SSE-NEXT:    [[ADD:%.*]] = add i64 undef, undef
75; SSE-NEXT:    store i64 [[ADD]], i64* undef, align 1
76; SSE-NEXT:    [[ARRAYIDX2_1:%.*]] = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 5
77; SSE-NEXT:    [[AND_1:%.*]] = shl i64 undef, 2
78; SSE-NEXT:    [[SHL_1:%.*]] = and i64 [[AND_1]], 20
79; SSE-NEXT:    [[SHR_1:%.*]] = lshr i64 undef, 6
80; SSE-NEXT:    [[ADD_1:%.*]] = add nuw nsw i64 [[SHL]], [[SHR_1]]
81; SSE-NEXT:    [[ARRAYIDX2_2:%.*]] = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 4
82; SSE-NEXT:    [[SHR_2:%.*]] = lshr i64 undef, 6
83; SSE-NEXT:    [[ADD_2:%.*]] = add nuw nsw i64 [[SHL_1]], [[SHR_2]]
84; SSE-NEXT:    [[AND_4:%.*]] = shl i64 [[ADD]], 2
85; SSE-NEXT:    [[SHL_4:%.*]] = and i64 [[AND_4]], 20
86; SSE-NEXT:    [[ARRAYIDX2_5:%.*]] = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 1
87; SSE-NEXT:    store i64 [[ADD_1]], i64* [[ARRAYIDX2_5]], align 1
88; SSE-NEXT:    [[AND_5:%.*]] = shl nuw nsw i64 [[ADD_1]], 2
89; SSE-NEXT:    [[SHL_5:%.*]] = and i64 [[AND_5]], 20
90; SSE-NEXT:    [[SHR_5:%.*]] = lshr i64 [[ADD_1]], 6
91; SSE-NEXT:    [[ADD_5:%.*]] = add nuw nsw i64 [[SHL_4]], [[SHR_5]]
92; SSE-NEXT:    store i64 [[ADD_5]], i64* [[ARRAYIDX2_1]], align 1
93; SSE-NEXT:    [[ARRAYIDX2_6:%.*]] = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 0
94; SSE-NEXT:    store i64 [[ADD_2]], i64* [[ARRAYIDX2_6]], align 1
95; SSE-NEXT:    [[SHR_6:%.*]] = lshr i64 [[ADD_2]], 6
96; SSE-NEXT:    [[ADD_6:%.*]] = add nuw nsw i64 [[SHL_5]], [[SHR_6]]
97; SSE-NEXT:    store i64 [[ADD_6]], i64* [[ARRAYIDX2_2]], align 1
98; SSE-NEXT:    ret void
99;
100; AVX-LABEL: @pr35497(
101; AVX-NEXT:  entry:
102; AVX-NEXT:    [[TMP0:%.*]] = load i64, i64* undef, align 1
103; AVX-NEXT:    [[ADD:%.*]] = add i64 undef, undef
104; AVX-NEXT:    store i64 [[ADD]], i64* undef, align 1
105; AVX-NEXT:    [[ARRAYIDX2_1:%.*]] = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 5
106; AVX-NEXT:    [[TMP1:%.*]] = insertelement <2 x i64> undef, i64 [[TMP0]], i32 1
107; AVX-NEXT:    [[TMP2:%.*]] = shl <2 x i64> [[TMP1]], <i64 2, i64 2>
108; AVX-NEXT:    [[TMP3:%.*]] = and <2 x i64> [[TMP2]], <i64 20, i64 20>
109; AVX-NEXT:    [[ARRAYIDX2_2:%.*]] = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 4
110; AVX-NEXT:    [[TMP4:%.*]] = add nuw nsw <2 x i64> [[TMP3]], zeroinitializer
111; AVX-NEXT:    [[ARRAYIDX2_5:%.*]] = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 1
112; AVX-NEXT:    [[TMP5:%.*]] = extractelement <2 x i64> [[TMP4]], i32 1
113; AVX-NEXT:    [[TMP6:%.*]] = insertelement <2 x i64> undef, i64 [[TMP5]], i32 0
114; AVX-NEXT:    [[TMP7:%.*]] = insertelement <2 x i64> [[TMP6]], i64 [[ADD]], i32 1
115; AVX-NEXT:    [[TMP8:%.*]] = shl <2 x i64> [[TMP7]], <i64 2, i64 2>
116; AVX-NEXT:    [[TMP9:%.*]] = and <2 x i64> [[TMP8]], <i64 20, i64 20>
117; AVX-NEXT:    [[ARRAYIDX2_6:%.*]] = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 0
118; AVX-NEXT:    [[TMP10:%.*]] = bitcast i64* [[ARRAYIDX2_6]] to <2 x i64>*
119; AVX-NEXT:    store <2 x i64> [[TMP4]], <2 x i64>* [[TMP10]], align 1
120; AVX-NEXT:    [[TMP11:%.*]] = extractelement <2 x i64> [[TMP4]], i32 0
121; AVX-NEXT:    [[TMP12:%.*]] = insertelement <2 x i64> undef, i64 [[TMP11]], i32 0
122; AVX-NEXT:    [[TMP13:%.*]] = insertelement <2 x i64> [[TMP12]], i64 [[TMP5]], i32 1
123; AVX-NEXT:    [[TMP14:%.*]] = lshr <2 x i64> [[TMP13]], <i64 6, i64 6>
124; AVX-NEXT:    [[TMP15:%.*]] = add nuw nsw <2 x i64> [[TMP9]], [[TMP14]]
125; AVX-NEXT:    [[TMP16:%.*]] = bitcast i64* [[ARRAYIDX2_2]] to <2 x i64>*
126; AVX-NEXT:    store <2 x i64> [[TMP15]], <2 x i64>* [[TMP16]], align 1
127; AVX-NEXT:    ret void
128;
129entry:
130  %0 = load i64, i64* undef, align 1
131  %and = shl i64 %0, 2
132  %shl = and i64 %and, 20
133  %add = add i64 undef, undef
134  store i64 %add, i64* undef, align 1
135  %arrayidx2.1 = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 5
136  %and.1 = shl i64 undef, 2
137  %shl.1 = and i64 %and.1, 20
138  %shr.1 = lshr i64 undef, 6
139  %add.1 = add nuw nsw i64 %shl, %shr.1
140  %arrayidx2.2 = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 4
141  %shr.2 = lshr i64 undef, 6
142  %add.2 = add nuw nsw i64 %shl.1, %shr.2
143  %and.4 = shl i64 %add, 2
144  %shl.4 = and i64 %and.4, 20
145  %arrayidx2.5 = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 1
146  store i64 %add.1, i64* %arrayidx2.5, align 1
147  %and.5 = shl nuw nsw i64 %add.1, 2
148  %shl.5 = and i64 %and.5, 20
149  %shr.5 = lshr i64 %add.1, 6
150  %add.5 = add nuw nsw i64 %shl.4, %shr.5
151  store i64 %add.5, i64* %arrayidx2.1, align 1
152  %arrayidx2.6 = getelementptr inbounds [0 x i64], [0 x i64]* undef, i64 0, i64 0
153  store i64 %add.2, i64* %arrayidx2.6, align 1
154  %shr.6 = lshr i64 %add.2, 6
155  %add.6 = add nuw nsw i64 %shl.5, %shr.6
156  store i64 %add.6, i64* %arrayidx2.2, align 1
157  ret void
158}
159