1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2; RUN:  opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux -mattr=+sse2     | FileCheck %s --check-prefixes=CHECK,SSE
3; RUN:  opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux -mattr=+avx      | FileCheck %s --check-prefixes=CHECK,AVX
4; RUN:  opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux -mattr=+avx2     | FileCheck %s --check-prefixes=CHECK,AVX2
5; RUN:  opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux -mattr=+avx512f  | FileCheck %s --check-prefixes=CHECK,AVX512F
6; RUN:  opt < %s -slp-vectorizer -instcombine -S -mtriple=x86_64-unknown-linux -mattr=+avx512vl | FileCheck %s --check-prefixes=CHECK,AVX512VL
7
8define void @gather_load(i32* noalias nocapture %0, i32* noalias nocapture readonly %1) {
9; CHECK-LABEL: @gather_load(
10; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i32, i32* [[TMP1:%.*]], i64 1
11; CHECK-NEXT:    [[TMP4:%.*]] = load i32, i32* [[TMP1]], align 4, !tbaa [[TBAA0:![0-9]+]]
12; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 11
13; CHECK-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4, !tbaa [[TBAA0]]
14; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 4
15; CHECK-NEXT:    [[TMP8:%.*]] = load i32, i32* [[TMP7]], align 4, !tbaa [[TBAA0]]
16; CHECK-NEXT:    [[TMP9:%.*]] = load i32, i32* [[TMP3]], align 4, !tbaa [[TBAA0]]
17; CHECK-NEXT:    [[TMP10:%.*]] = insertelement <4 x i32> undef, i32 [[TMP4]], i64 0
18; CHECK-NEXT:    [[TMP11:%.*]] = insertelement <4 x i32> [[TMP10]], i32 [[TMP6]], i64 1
19; CHECK-NEXT:    [[TMP12:%.*]] = insertelement <4 x i32> [[TMP11]], i32 [[TMP8]], i64 2
20; CHECK-NEXT:    [[TMP13:%.*]] = insertelement <4 x i32> [[TMP12]], i32 [[TMP9]], i64 3
21; CHECK-NEXT:    [[TMP14:%.*]] = add nsw <4 x i32> [[TMP13]], <i32 1, i32 2, i32 3, i32 4>
22; CHECK-NEXT:    [[TMP15:%.*]] = bitcast i32* [[TMP0:%.*]] to <4 x i32>*
23; CHECK-NEXT:    store <4 x i32> [[TMP14]], <4 x i32>* [[TMP15]], align 4, !tbaa [[TBAA0]]
24; CHECK-NEXT:    ret void
25;
26  %3 = getelementptr inbounds i32, i32* %1, i64 1
27  %4 = load i32, i32* %1, align 4, !tbaa !2
28  %5 = getelementptr inbounds i32, i32* %0, i64 1
29  %6 = getelementptr inbounds i32, i32* %1, i64 11
30  %7 = load i32, i32* %6, align 4, !tbaa !2
31  %8 = getelementptr inbounds i32, i32* %0, i64 2
32  %9 = getelementptr inbounds i32, i32* %1, i64 4
33  %10 = load i32, i32* %9, align 4, !tbaa !2
34  %11 = getelementptr inbounds i32, i32* %0, i64 3
35  %12 = load i32, i32* %3, align 4, !tbaa !2
36  %13 = insertelement <4 x i32> undef, i32 %4, i32 0
37  %14 = insertelement <4 x i32> %13, i32 %7, i32 1
38  %15 = insertelement <4 x i32> %14, i32 %10, i32 2
39  %16 = insertelement <4 x i32> %15, i32 %12, i32 3
40  %17 = add nsw <4 x i32> %16, <i32 1, i32 2, i32 3, i32 4>
41  %18 = bitcast i32* %0 to <4 x i32>*
42  store <4 x i32> %17, <4 x i32>* %18, align 4, !tbaa !2
43  ret void
44}
45
46define void @gather_load_2(i32* noalias nocapture %0, i32* noalias nocapture readonly %1) {
47; SSE-LABEL: @gather_load_2(
48; SSE-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i32, i32* [[TMP1:%.*]], i64 1
49; SSE-NEXT:    [[TMP4:%.*]] = load i32, i32* [[TMP3]], align 4, !tbaa [[TBAA0]]
50; SSE-NEXT:    [[TMP5:%.*]] = add nsw i32 [[TMP4]], 1
51; SSE-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TMP0:%.*]], i64 1
52; SSE-NEXT:    store i32 [[TMP5]], i32* [[TMP0]], align 4, !tbaa [[TBAA0]]
53; SSE-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 10
54; SSE-NEXT:    [[TMP8:%.*]] = load i32, i32* [[TMP7]], align 4, !tbaa [[TBAA0]]
55; SSE-NEXT:    [[TMP9:%.*]] = add nsw i32 [[TMP8]], 2
56; SSE-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 2
57; SSE-NEXT:    store i32 [[TMP9]], i32* [[TMP6]], align 4, !tbaa [[TBAA0]]
58; SSE-NEXT:    [[TMP11:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 3
59; SSE-NEXT:    [[TMP12:%.*]] = load i32, i32* [[TMP11]], align 4, !tbaa [[TBAA0]]
60; SSE-NEXT:    [[TMP13:%.*]] = add nsw i32 [[TMP12]], 3
61; SSE-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 3
62; SSE-NEXT:    store i32 [[TMP13]], i32* [[TMP10]], align 4, !tbaa [[TBAA0]]
63; SSE-NEXT:    [[TMP15:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 5
64; SSE-NEXT:    [[TMP16:%.*]] = load i32, i32* [[TMP15]], align 4, !tbaa [[TBAA0]]
65; SSE-NEXT:    [[TMP17:%.*]] = add nsw i32 [[TMP16]], 4
66; SSE-NEXT:    store i32 [[TMP17]], i32* [[TMP14]], align 4, !tbaa [[TBAA0]]
67; SSE-NEXT:    ret void
68;
69; AVX-LABEL: @gather_load_2(
70; AVX-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i32, i32* [[TMP1:%.*]], i64 1
71; AVX-NEXT:    [[TMP4:%.*]] = load i32, i32* [[TMP3]], align 4, !tbaa [[TBAA0]]
72; AVX-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 10
73; AVX-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4, !tbaa [[TBAA0]]
74; AVX-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 3
75; AVX-NEXT:    [[TMP8:%.*]] = load i32, i32* [[TMP7]], align 4, !tbaa [[TBAA0]]
76; AVX-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 5
77; AVX-NEXT:    [[TMP10:%.*]] = load i32, i32* [[TMP9]], align 4, !tbaa [[TBAA0]]
78; AVX-NEXT:    [[TMP11:%.*]] = insertelement <4 x i32> poison, i32 [[TMP4]], i64 0
79; AVX-NEXT:    [[TMP12:%.*]] = insertelement <4 x i32> [[TMP11]], i32 [[TMP6]], i64 1
80; AVX-NEXT:    [[TMP13:%.*]] = insertelement <4 x i32> [[TMP12]], i32 [[TMP8]], i64 2
81; AVX-NEXT:    [[TMP14:%.*]] = insertelement <4 x i32> [[TMP13]], i32 [[TMP10]], i64 3
82; AVX-NEXT:    [[TMP15:%.*]] = add nsw <4 x i32> [[TMP14]], <i32 1, i32 2, i32 3, i32 4>
83; AVX-NEXT:    [[TMP16:%.*]] = bitcast i32* [[TMP0:%.*]] to <4 x i32>*
84; AVX-NEXT:    store <4 x i32> [[TMP15]], <4 x i32>* [[TMP16]], align 4, !tbaa [[TBAA0]]
85; AVX-NEXT:    ret void
86;
87; AVX2-LABEL: @gather_load_2(
88; AVX2-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i32, i32* [[TMP1:%.*]], i64 1
89; AVX2-NEXT:    [[TMP4:%.*]] = load i32, i32* [[TMP3]], align 4, !tbaa [[TBAA0]]
90; AVX2-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 10
91; AVX2-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4, !tbaa [[TBAA0]]
92; AVX2-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 3
93; AVX2-NEXT:    [[TMP8:%.*]] = load i32, i32* [[TMP7]], align 4, !tbaa [[TBAA0]]
94; AVX2-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 5
95; AVX2-NEXT:    [[TMP10:%.*]] = load i32, i32* [[TMP9]], align 4, !tbaa [[TBAA0]]
96; AVX2-NEXT:    [[TMP11:%.*]] = insertelement <4 x i32> poison, i32 [[TMP4]], i64 0
97; AVX2-NEXT:    [[TMP12:%.*]] = insertelement <4 x i32> [[TMP11]], i32 [[TMP6]], i64 1
98; AVX2-NEXT:    [[TMP13:%.*]] = insertelement <4 x i32> [[TMP12]], i32 [[TMP8]], i64 2
99; AVX2-NEXT:    [[TMP14:%.*]] = insertelement <4 x i32> [[TMP13]], i32 [[TMP10]], i64 3
100; AVX2-NEXT:    [[TMP15:%.*]] = add nsw <4 x i32> [[TMP14]], <i32 1, i32 2, i32 3, i32 4>
101; AVX2-NEXT:    [[TMP16:%.*]] = bitcast i32* [[TMP0:%.*]] to <4 x i32>*
102; AVX2-NEXT:    store <4 x i32> [[TMP15]], <4 x i32>* [[TMP16]], align 4, !tbaa [[TBAA0]]
103; AVX2-NEXT:    ret void
104;
105; AVX512F-LABEL: @gather_load_2(
106; AVX512F-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i32, i32* [[TMP1:%.*]], i64 1
107; AVX512F-NEXT:    [[TMP4:%.*]] = load i32, i32* [[TMP3]], align 4, !tbaa [[TBAA0]]
108; AVX512F-NEXT:    [[TMP5:%.*]] = add nsw i32 [[TMP4]], 1
109; AVX512F-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TMP0:%.*]], i64 1
110; AVX512F-NEXT:    store i32 [[TMP5]], i32* [[TMP0]], align 4, !tbaa [[TBAA0]]
111; AVX512F-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 10
112; AVX512F-NEXT:    [[TMP8:%.*]] = load i32, i32* [[TMP7]], align 4, !tbaa [[TBAA0]]
113; AVX512F-NEXT:    [[TMP9:%.*]] = add nsw i32 [[TMP8]], 2
114; AVX512F-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 2
115; AVX512F-NEXT:    store i32 [[TMP9]], i32* [[TMP6]], align 4, !tbaa [[TBAA0]]
116; AVX512F-NEXT:    [[TMP11:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 3
117; AVX512F-NEXT:    [[TMP12:%.*]] = load i32, i32* [[TMP11]], align 4, !tbaa [[TBAA0]]
118; AVX512F-NEXT:    [[TMP13:%.*]] = add nsw i32 [[TMP12]], 3
119; AVX512F-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 3
120; AVX512F-NEXT:    store i32 [[TMP13]], i32* [[TMP10]], align 4, !tbaa [[TBAA0]]
121; AVX512F-NEXT:    [[TMP15:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 5
122; AVX512F-NEXT:    [[TMP16:%.*]] = load i32, i32* [[TMP15]], align 4, !tbaa [[TBAA0]]
123; AVX512F-NEXT:    [[TMP17:%.*]] = add nsw i32 [[TMP16]], 4
124; AVX512F-NEXT:    store i32 [[TMP17]], i32* [[TMP14]], align 4, !tbaa [[TBAA0]]
125; AVX512F-NEXT:    ret void
126;
127; AVX512VL-LABEL: @gather_load_2(
128; AVX512VL-NEXT:    [[TMP3:%.*]] = insertelement <4 x i32*> poison, i32* [[TMP1:%.*]], i64 0
129; AVX512VL-NEXT:    [[SHUFFLE:%.*]] = shufflevector <4 x i32*> [[TMP3]], <4 x i32*> poison, <4 x i32> zeroinitializer
130; AVX512VL-NEXT:    [[TMP4:%.*]] = getelementptr i32, <4 x i32*> [[SHUFFLE]], <4 x i64> <i64 1, i64 10, i64 3, i64 5>
131; AVX512VL-NEXT:    [[TMP5:%.*]] = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> [[TMP4]], i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef), !tbaa [[TBAA0]]
132; AVX512VL-NEXT:    [[TMP6:%.*]] = add nsw <4 x i32> [[TMP5]], <i32 1, i32 2, i32 3, i32 4>
133; AVX512VL-NEXT:    [[TMP7:%.*]] = bitcast i32* [[TMP0:%.*]] to <4 x i32>*
134; AVX512VL-NEXT:    store <4 x i32> [[TMP6]], <4 x i32>* [[TMP7]], align 4, !tbaa [[TBAA0]]
135; AVX512VL-NEXT:    ret void
136;
137  %3 = getelementptr inbounds i32, i32* %1, i64 1
138  %4 = load i32, i32* %3, align 4, !tbaa !2
139  %5 = add nsw i32 %4, 1
140  %6 = getelementptr inbounds i32, i32* %0, i64 1
141  store i32 %5, i32* %0, align 4, !tbaa !2
142  %7 = getelementptr inbounds i32, i32* %1, i64 10
143  %8 = load i32, i32* %7, align 4, !tbaa !2
144  %9 = add nsw i32 %8, 2
145  %10 = getelementptr inbounds i32, i32* %0, i64 2
146  store i32 %9, i32* %6, align 4, !tbaa !2
147  %11 = getelementptr inbounds i32, i32* %1, i64 3
148  %12 = load i32, i32* %11, align 4, !tbaa !2
149  %13 = add nsw i32 %12, 3
150  %14 = getelementptr inbounds i32, i32* %0, i64 3
151  store i32 %13, i32* %10, align 4, !tbaa !2
152  %15 = getelementptr inbounds i32, i32* %1, i64 5
153  %16 = load i32, i32* %15, align 4, !tbaa !2
154  %17 = add nsw i32 %16, 4
155  store i32 %17, i32* %14, align 4, !tbaa !2
156  ret void
157}
158
159
160define void @gather_load_3(i32* noalias nocapture %0, i32* noalias nocapture readonly %1) {
161; SSE-LABEL: @gather_load_3(
162; SSE-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP1:%.*]], align 4, !tbaa [[TBAA0]]
163; SSE-NEXT:    [[TMP4:%.*]] = add i32 [[TMP3]], 1
164; SSE-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[TMP0:%.*]], i64 1
165; SSE-NEXT:    store i32 [[TMP4]], i32* [[TMP0]], align 4, !tbaa [[TBAA0]]
166; SSE-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 11
167; SSE-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP6]], align 4, !tbaa [[TBAA0]]
168; SSE-NEXT:    [[TMP8:%.*]] = add i32 [[TMP7]], 2
169; SSE-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 2
170; SSE-NEXT:    store i32 [[TMP8]], i32* [[TMP5]], align 4, !tbaa [[TBAA0]]
171; SSE-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 4
172; SSE-NEXT:    [[TMP11:%.*]] = load i32, i32* [[TMP10]], align 4, !tbaa [[TBAA0]]
173; SSE-NEXT:    [[TMP12:%.*]] = add i32 [[TMP11]], 3
174; SSE-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 3
175; SSE-NEXT:    store i32 [[TMP12]], i32* [[TMP9]], align 4, !tbaa [[TBAA0]]
176; SSE-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 15
177; SSE-NEXT:    [[TMP15:%.*]] = load i32, i32* [[TMP14]], align 4, !tbaa [[TBAA0]]
178; SSE-NEXT:    [[TMP16:%.*]] = add i32 [[TMP15]], 4
179; SSE-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 4
180; SSE-NEXT:    store i32 [[TMP16]], i32* [[TMP13]], align 4, !tbaa [[TBAA0]]
181; SSE-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 18
182; SSE-NEXT:    [[TMP19:%.*]] = load i32, i32* [[TMP18]], align 4, !tbaa [[TBAA0]]
183; SSE-NEXT:    [[TMP20:%.*]] = add i32 [[TMP19]], 1
184; SSE-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 5
185; SSE-NEXT:    store i32 [[TMP20]], i32* [[TMP17]], align 4, !tbaa [[TBAA0]]
186; SSE-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 9
187; SSE-NEXT:    [[TMP23:%.*]] = load i32, i32* [[TMP22]], align 4, !tbaa [[TBAA0]]
188; SSE-NEXT:    [[TMP24:%.*]] = add i32 [[TMP23]], 2
189; SSE-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 6
190; SSE-NEXT:    store i32 [[TMP24]], i32* [[TMP21]], align 4, !tbaa [[TBAA0]]
191; SSE-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 6
192; SSE-NEXT:    [[TMP27:%.*]] = load i32, i32* [[TMP26]], align 4, !tbaa [[TBAA0]]
193; SSE-NEXT:    [[TMP28:%.*]] = add i32 [[TMP27]], 3
194; SSE-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 7
195; SSE-NEXT:    store i32 [[TMP28]], i32* [[TMP25]], align 4, !tbaa [[TBAA0]]
196; SSE-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 21
197; SSE-NEXT:    [[TMP31:%.*]] = load i32, i32* [[TMP30]], align 4, !tbaa [[TBAA0]]
198; SSE-NEXT:    [[TMP32:%.*]] = add i32 [[TMP31]], 4
199; SSE-NEXT:    store i32 [[TMP32]], i32* [[TMP29]], align 4, !tbaa [[TBAA0]]
200; SSE-NEXT:    ret void
201;
202; AVX-LABEL: @gather_load_3(
203; AVX-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP1:%.*]], align 4, !tbaa [[TBAA0]]
204; AVX-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 11
205; AVX-NEXT:    [[TMP5:%.*]] = load i32, i32* [[TMP4]], align 4, !tbaa [[TBAA0]]
206; AVX-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 4
207; AVX-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP6]], align 4, !tbaa [[TBAA0]]
208; AVX-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 15
209; AVX-NEXT:    [[TMP9:%.*]] = load i32, i32* [[TMP8]], align 4, !tbaa [[TBAA0]]
210; AVX-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 18
211; AVX-NEXT:    [[TMP11:%.*]] = load i32, i32* [[TMP10]], align 4, !tbaa [[TBAA0]]
212; AVX-NEXT:    [[TMP12:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 9
213; AVX-NEXT:    [[TMP13:%.*]] = load i32, i32* [[TMP12]], align 4, !tbaa [[TBAA0]]
214; AVX-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 6
215; AVX-NEXT:    [[TMP15:%.*]] = load i32, i32* [[TMP14]], align 4, !tbaa [[TBAA0]]
216; AVX-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 21
217; AVX-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 4, !tbaa [[TBAA0]]
218; AVX-NEXT:    [[TMP18:%.*]] = insertelement <8 x i32> poison, i32 [[TMP3]], i64 0
219; AVX-NEXT:    [[TMP19:%.*]] = insertelement <8 x i32> [[TMP18]], i32 [[TMP5]], i64 1
220; AVX-NEXT:    [[TMP20:%.*]] = insertelement <8 x i32> [[TMP19]], i32 [[TMP7]], i64 2
221; AVX-NEXT:    [[TMP21:%.*]] = insertelement <8 x i32> [[TMP20]], i32 [[TMP9]], i64 3
222; AVX-NEXT:    [[TMP22:%.*]] = insertelement <8 x i32> [[TMP21]], i32 [[TMP11]], i64 4
223; AVX-NEXT:    [[TMP23:%.*]] = insertelement <8 x i32> [[TMP22]], i32 [[TMP13]], i64 5
224; AVX-NEXT:    [[TMP24:%.*]] = insertelement <8 x i32> [[TMP23]], i32 [[TMP15]], i64 6
225; AVX-NEXT:    [[TMP25:%.*]] = insertelement <8 x i32> [[TMP24]], i32 [[TMP17]], i64 7
226; AVX-NEXT:    [[TMP26:%.*]] = add <8 x i32> [[TMP25]], <i32 1, i32 2, i32 3, i32 4, i32 1, i32 2, i32 3, i32 4>
227; AVX-NEXT:    [[TMP27:%.*]] = bitcast i32* [[TMP0:%.*]] to <8 x i32>*
228; AVX-NEXT:    store <8 x i32> [[TMP26]], <8 x i32>* [[TMP27]], align 4, !tbaa [[TBAA0]]
229; AVX-NEXT:    ret void
230;
231; AVX2-LABEL: @gather_load_3(
232; AVX2-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP1:%.*]], align 4, !tbaa [[TBAA0]]
233; AVX2-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 11
234; AVX2-NEXT:    [[TMP5:%.*]] = load i32, i32* [[TMP4]], align 4, !tbaa [[TBAA0]]
235; AVX2-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 4
236; AVX2-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP6]], align 4, !tbaa [[TBAA0]]
237; AVX2-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 15
238; AVX2-NEXT:    [[TMP9:%.*]] = load i32, i32* [[TMP8]], align 4, !tbaa [[TBAA0]]
239; AVX2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 18
240; AVX2-NEXT:    [[TMP11:%.*]] = load i32, i32* [[TMP10]], align 4, !tbaa [[TBAA0]]
241; AVX2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 9
242; AVX2-NEXT:    [[TMP13:%.*]] = load i32, i32* [[TMP12]], align 4, !tbaa [[TBAA0]]
243; AVX2-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 6
244; AVX2-NEXT:    [[TMP15:%.*]] = load i32, i32* [[TMP14]], align 4, !tbaa [[TBAA0]]
245; AVX2-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 21
246; AVX2-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 4, !tbaa [[TBAA0]]
247; AVX2-NEXT:    [[TMP18:%.*]] = insertelement <8 x i32> poison, i32 [[TMP3]], i64 0
248; AVX2-NEXT:    [[TMP19:%.*]] = insertelement <8 x i32> [[TMP18]], i32 [[TMP5]], i64 1
249; AVX2-NEXT:    [[TMP20:%.*]] = insertelement <8 x i32> [[TMP19]], i32 [[TMP7]], i64 2
250; AVX2-NEXT:    [[TMP21:%.*]] = insertelement <8 x i32> [[TMP20]], i32 [[TMP9]], i64 3
251; AVX2-NEXT:    [[TMP22:%.*]] = insertelement <8 x i32> [[TMP21]], i32 [[TMP11]], i64 4
252; AVX2-NEXT:    [[TMP23:%.*]] = insertelement <8 x i32> [[TMP22]], i32 [[TMP13]], i64 5
253; AVX2-NEXT:    [[TMP24:%.*]] = insertelement <8 x i32> [[TMP23]], i32 [[TMP15]], i64 6
254; AVX2-NEXT:    [[TMP25:%.*]] = insertelement <8 x i32> [[TMP24]], i32 [[TMP17]], i64 7
255; AVX2-NEXT:    [[TMP26:%.*]] = add <8 x i32> [[TMP25]], <i32 1, i32 2, i32 3, i32 4, i32 1, i32 2, i32 3, i32 4>
256; AVX2-NEXT:    [[TMP27:%.*]] = bitcast i32* [[TMP0:%.*]] to <8 x i32>*
257; AVX2-NEXT:    store <8 x i32> [[TMP26]], <8 x i32>* [[TMP27]], align 4, !tbaa [[TBAA0]]
258; AVX2-NEXT:    ret void
259;
260; AVX512F-LABEL: @gather_load_3(
261; AVX512F-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP1:%.*]], align 4, !tbaa [[TBAA0]]
262; AVX512F-NEXT:    [[TMP4:%.*]] = add i32 [[TMP3]], 1
263; AVX512F-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[TMP0:%.*]], i64 1
264; AVX512F-NEXT:    store i32 [[TMP4]], i32* [[TMP0]], align 4, !tbaa [[TBAA0]]
265; AVX512F-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 11
266; AVX512F-NEXT:    [[TMP7:%.*]] = load i32, i32* [[TMP6]], align 4, !tbaa [[TBAA0]]
267; AVX512F-NEXT:    [[TMP8:%.*]] = add i32 [[TMP7]], 2
268; AVX512F-NEXT:    [[TMP9:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 2
269; AVX512F-NEXT:    store i32 [[TMP8]], i32* [[TMP5]], align 4, !tbaa [[TBAA0]]
270; AVX512F-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 4
271; AVX512F-NEXT:    [[TMP11:%.*]] = load i32, i32* [[TMP10]], align 4, !tbaa [[TBAA0]]
272; AVX512F-NEXT:    [[TMP12:%.*]] = add i32 [[TMP11]], 3
273; AVX512F-NEXT:    [[TMP13:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 3
274; AVX512F-NEXT:    store i32 [[TMP12]], i32* [[TMP9]], align 4, !tbaa [[TBAA0]]
275; AVX512F-NEXT:    [[TMP14:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 15
276; AVX512F-NEXT:    [[TMP15:%.*]] = load i32, i32* [[TMP14]], align 4, !tbaa [[TBAA0]]
277; AVX512F-NEXT:    [[TMP16:%.*]] = add i32 [[TMP15]], 4
278; AVX512F-NEXT:    [[TMP17:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 4
279; AVX512F-NEXT:    store i32 [[TMP16]], i32* [[TMP13]], align 4, !tbaa [[TBAA0]]
280; AVX512F-NEXT:    [[TMP18:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 18
281; AVX512F-NEXT:    [[TMP19:%.*]] = load i32, i32* [[TMP18]], align 4, !tbaa [[TBAA0]]
282; AVX512F-NEXT:    [[TMP20:%.*]] = add i32 [[TMP19]], 1
283; AVX512F-NEXT:    [[TMP21:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 5
284; AVX512F-NEXT:    store i32 [[TMP20]], i32* [[TMP17]], align 4, !tbaa [[TBAA0]]
285; AVX512F-NEXT:    [[TMP22:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 9
286; AVX512F-NEXT:    [[TMP23:%.*]] = load i32, i32* [[TMP22]], align 4, !tbaa [[TBAA0]]
287; AVX512F-NEXT:    [[TMP24:%.*]] = add i32 [[TMP23]], 2
288; AVX512F-NEXT:    [[TMP25:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 6
289; AVX512F-NEXT:    store i32 [[TMP24]], i32* [[TMP21]], align 4, !tbaa [[TBAA0]]
290; AVX512F-NEXT:    [[TMP26:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 6
291; AVX512F-NEXT:    [[TMP27:%.*]] = load i32, i32* [[TMP26]], align 4, !tbaa [[TBAA0]]
292; AVX512F-NEXT:    [[TMP28:%.*]] = add i32 [[TMP27]], 3
293; AVX512F-NEXT:    [[TMP29:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 7
294; AVX512F-NEXT:    store i32 [[TMP28]], i32* [[TMP25]], align 4, !tbaa [[TBAA0]]
295; AVX512F-NEXT:    [[TMP30:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 21
296; AVX512F-NEXT:    [[TMP31:%.*]] = load i32, i32* [[TMP30]], align 4, !tbaa [[TBAA0]]
297; AVX512F-NEXT:    [[TMP32:%.*]] = add i32 [[TMP31]], 4
298; AVX512F-NEXT:    store i32 [[TMP32]], i32* [[TMP29]], align 4, !tbaa [[TBAA0]]
299; AVX512F-NEXT:    ret void
300;
301; AVX512VL-LABEL: @gather_load_3(
302; AVX512VL-NEXT:    [[TMP3:%.*]] = load i32, i32* [[TMP1:%.*]], align 4, !tbaa [[TBAA0]]
303; AVX512VL-NEXT:    [[TMP4:%.*]] = add i32 [[TMP3]], 1
304; AVX512VL-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i32, i32* [[TMP0:%.*]], i64 1
305; AVX512VL-NEXT:    store i32 [[TMP4]], i32* [[TMP0]], align 4, !tbaa [[TBAA0]]
306; AVX512VL-NEXT:    [[TMP6:%.*]] = insertelement <4 x i32*> poison, i32* [[TMP1]], i64 0
307; AVX512VL-NEXT:    [[SHUFFLE:%.*]] = shufflevector <4 x i32*> [[TMP6]], <4 x i32*> poison, <4 x i32> zeroinitializer
308; AVX512VL-NEXT:    [[TMP7:%.*]] = getelementptr i32, <4 x i32*> [[SHUFFLE]], <4 x i64> <i64 11, i64 4, i64 15, i64 18>
309; AVX512VL-NEXT:    [[TMP8:%.*]] = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> [[TMP7]], i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef), !tbaa [[TBAA0]]
310; AVX512VL-NEXT:    [[TMP9:%.*]] = add <4 x i32> [[TMP8]], <i32 2, i32 3, i32 4, i32 1>
311; AVX512VL-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 5
312; AVX512VL-NEXT:    [[TMP11:%.*]] = bitcast i32* [[TMP5]] to <4 x i32>*
313; AVX512VL-NEXT:    store <4 x i32> [[TMP9]], <4 x i32>* [[TMP11]], align 4, !tbaa [[TBAA0]]
314; AVX512VL-NEXT:    [[TMP12:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 9
315; AVX512VL-NEXT:    [[TMP13:%.*]] = load i32, i32* [[TMP12]], align 4, !tbaa [[TBAA0]]
316; AVX512VL-NEXT:    [[TMP14:%.*]] = add i32 [[TMP13]], 2
317; AVX512VL-NEXT:    [[TMP15:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 6
318; AVX512VL-NEXT:    store i32 [[TMP14]], i32* [[TMP10]], align 4, !tbaa [[TBAA0]]
319; AVX512VL-NEXT:    [[TMP16:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 6
320; AVX512VL-NEXT:    [[TMP17:%.*]] = load i32, i32* [[TMP16]], align 4, !tbaa [[TBAA0]]
321; AVX512VL-NEXT:    [[TMP18:%.*]] = add i32 [[TMP17]], 3
322; AVX512VL-NEXT:    [[TMP19:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 7
323; AVX512VL-NEXT:    store i32 [[TMP18]], i32* [[TMP15]], align 4, !tbaa [[TBAA0]]
324; AVX512VL-NEXT:    [[TMP20:%.*]] = getelementptr inbounds i32, i32* [[TMP1]], i64 21
325; AVX512VL-NEXT:    [[TMP21:%.*]] = load i32, i32* [[TMP20]], align 4, !tbaa [[TBAA0]]
326; AVX512VL-NEXT:    [[TMP22:%.*]] = add i32 [[TMP21]], 4
327; AVX512VL-NEXT:    store i32 [[TMP22]], i32* [[TMP19]], align 4, !tbaa [[TBAA0]]
328; AVX512VL-NEXT:    ret void
329;
330  %3 = load i32, i32* %1, align 4, !tbaa !2
331  %4 = add i32 %3, 1
332  %5 = getelementptr inbounds i32, i32* %0, i64 1
333  store i32 %4, i32* %0, align 4, !tbaa !2
334  %6 = getelementptr inbounds i32, i32* %1, i64 11
335  %7 = load i32, i32* %6, align 4, !tbaa !2
336  %8 = add i32 %7, 2
337  %9 = getelementptr inbounds i32, i32* %0, i64 2
338  store i32 %8, i32* %5, align 4, !tbaa !2
339  %10 = getelementptr inbounds i32, i32* %1, i64 4
340  %11 = load i32, i32* %10, align 4, !tbaa !2
341  %12 = add i32 %11, 3
342  %13 = getelementptr inbounds i32, i32* %0, i64 3
343  store i32 %12, i32* %9, align 4, !tbaa !2
344  %14 = getelementptr inbounds i32, i32* %1, i64 15
345  %15 = load i32, i32* %14, align 4, !tbaa !2
346  %16 = add i32 %15, 4
347  %17 = getelementptr inbounds i32, i32* %0, i64 4
348  store i32 %16, i32* %13, align 4, !tbaa !2
349  %18 = getelementptr inbounds i32, i32* %1, i64 18
350  %19 = load i32, i32* %18, align 4, !tbaa !2
351  %20 = add i32 %19, 1
352  %21 = getelementptr inbounds i32, i32* %0, i64 5
353  store i32 %20, i32* %17, align 4, !tbaa !2
354  %22 = getelementptr inbounds i32, i32* %1, i64 9
355  %23 = load i32, i32* %22, align 4, !tbaa !2
356  %24 = add i32 %23, 2
357  %25 = getelementptr inbounds i32, i32* %0, i64 6
358  store i32 %24, i32* %21, align 4, !tbaa !2
359  %26 = getelementptr inbounds i32, i32* %1, i64 6
360  %27 = load i32, i32* %26, align 4, !tbaa !2
361  %28 = add i32 %27, 3
362  %29 = getelementptr inbounds i32, i32* %0, i64 7
363  store i32 %28, i32* %25, align 4, !tbaa !2
364  %30 = getelementptr inbounds i32, i32* %1, i64 21
365  %31 = load i32, i32* %30, align 4, !tbaa !2
366  %32 = add i32 %31, 4
367  store i32 %32, i32* %29, align 4, !tbaa !2
368  ret void
369}
370
371define void @gather_load_4(i32* noalias nocapture %t0, i32* noalias nocapture readonly %t1) {
372; SSE-LABEL: @gather_load_4(
373; SSE-NEXT:    [[T5:%.*]] = getelementptr inbounds i32, i32* [[T0:%.*]], i64 1
374; SSE-NEXT:    [[T6:%.*]] = getelementptr inbounds i32, i32* [[T1:%.*]], i64 11
375; SSE-NEXT:    [[T9:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 2
376; SSE-NEXT:    [[T10:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 4
377; SSE-NEXT:    [[T13:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 3
378; SSE-NEXT:    [[T14:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 15
379; SSE-NEXT:    [[T17:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 4
380; SSE-NEXT:    [[T18:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 18
381; SSE-NEXT:    [[T21:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 5
382; SSE-NEXT:    [[T22:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 9
383; SSE-NEXT:    [[T25:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 6
384; SSE-NEXT:    [[T26:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 6
385; SSE-NEXT:    [[T29:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 7
386; SSE-NEXT:    [[T30:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 21
387; SSE-NEXT:    [[T3:%.*]] = load i32, i32* [[T1]], align 4, !tbaa [[TBAA0]]
388; SSE-NEXT:    [[T7:%.*]] = load i32, i32* [[T6]], align 4, !tbaa [[TBAA0]]
389; SSE-NEXT:    [[T11:%.*]] = load i32, i32* [[T10]], align 4, !tbaa [[TBAA0]]
390; SSE-NEXT:    [[T15:%.*]] = load i32, i32* [[T14]], align 4, !tbaa [[TBAA0]]
391; SSE-NEXT:    [[T19:%.*]] = load i32, i32* [[T18]], align 4, !tbaa [[TBAA0]]
392; SSE-NEXT:    [[T23:%.*]] = load i32, i32* [[T22]], align 4, !tbaa [[TBAA0]]
393; SSE-NEXT:    [[T27:%.*]] = load i32, i32* [[T26]], align 4, !tbaa [[TBAA0]]
394; SSE-NEXT:    [[T31:%.*]] = load i32, i32* [[T30]], align 4, !tbaa [[TBAA0]]
395; SSE-NEXT:    [[T4:%.*]] = add i32 [[T3]], 1
396; SSE-NEXT:    [[T8:%.*]] = add i32 [[T7]], 2
397; SSE-NEXT:    [[T12:%.*]] = add i32 [[T11]], 3
398; SSE-NEXT:    [[T16:%.*]] = add i32 [[T15]], 4
399; SSE-NEXT:    [[T20:%.*]] = add i32 [[T19]], 1
400; SSE-NEXT:    [[T24:%.*]] = add i32 [[T23]], 2
401; SSE-NEXT:    [[T28:%.*]] = add i32 [[T27]], 3
402; SSE-NEXT:    [[T32:%.*]] = add i32 [[T31]], 4
403; SSE-NEXT:    store i32 [[T4]], i32* [[T0]], align 4, !tbaa [[TBAA0]]
404; SSE-NEXT:    store i32 [[T8]], i32* [[T5]], align 4, !tbaa [[TBAA0]]
405; SSE-NEXT:    store i32 [[T12]], i32* [[T9]], align 4, !tbaa [[TBAA0]]
406; SSE-NEXT:    store i32 [[T16]], i32* [[T13]], align 4, !tbaa [[TBAA0]]
407; SSE-NEXT:    store i32 [[T20]], i32* [[T17]], align 4, !tbaa [[TBAA0]]
408; SSE-NEXT:    store i32 [[T24]], i32* [[T21]], align 4, !tbaa [[TBAA0]]
409; SSE-NEXT:    store i32 [[T28]], i32* [[T25]], align 4, !tbaa [[TBAA0]]
410; SSE-NEXT:    store i32 [[T32]], i32* [[T29]], align 4, !tbaa [[TBAA0]]
411; SSE-NEXT:    ret void
412;
413; AVX-LABEL: @gather_load_4(
414; AVX-NEXT:    [[T6:%.*]] = getelementptr inbounds i32, i32* [[T1:%.*]], i64 11
415; AVX-NEXT:    [[T10:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 4
416; AVX-NEXT:    [[T14:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 15
417; AVX-NEXT:    [[T18:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 18
418; AVX-NEXT:    [[T22:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 9
419; AVX-NEXT:    [[T26:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 6
420; AVX-NEXT:    [[T30:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 21
421; AVX-NEXT:    [[T3:%.*]] = load i32, i32* [[T1]], align 4, !tbaa [[TBAA0]]
422; AVX-NEXT:    [[T7:%.*]] = load i32, i32* [[T6]], align 4, !tbaa [[TBAA0]]
423; AVX-NEXT:    [[T11:%.*]] = load i32, i32* [[T10]], align 4, !tbaa [[TBAA0]]
424; AVX-NEXT:    [[T15:%.*]] = load i32, i32* [[T14]], align 4, !tbaa [[TBAA0]]
425; AVX-NEXT:    [[T19:%.*]] = load i32, i32* [[T18]], align 4, !tbaa [[TBAA0]]
426; AVX-NEXT:    [[T23:%.*]] = load i32, i32* [[T22]], align 4, !tbaa [[TBAA0]]
427; AVX-NEXT:    [[T27:%.*]] = load i32, i32* [[T26]], align 4, !tbaa [[TBAA0]]
428; AVX-NEXT:    [[T31:%.*]] = load i32, i32* [[T30]], align 4, !tbaa [[TBAA0]]
429; AVX-NEXT:    [[TMP1:%.*]] = insertelement <8 x i32> poison, i32 [[T3]], i64 0
430; AVX-NEXT:    [[TMP2:%.*]] = insertelement <8 x i32> [[TMP1]], i32 [[T7]], i64 1
431; AVX-NEXT:    [[TMP3:%.*]] = insertelement <8 x i32> [[TMP2]], i32 [[T11]], i64 2
432; AVX-NEXT:    [[TMP4:%.*]] = insertelement <8 x i32> [[TMP3]], i32 [[T15]], i64 3
433; AVX-NEXT:    [[TMP5:%.*]] = insertelement <8 x i32> [[TMP4]], i32 [[T19]], i64 4
434; AVX-NEXT:    [[TMP6:%.*]] = insertelement <8 x i32> [[TMP5]], i32 [[T23]], i64 5
435; AVX-NEXT:    [[TMP7:%.*]] = insertelement <8 x i32> [[TMP6]], i32 [[T27]], i64 6
436; AVX-NEXT:    [[TMP8:%.*]] = insertelement <8 x i32> [[TMP7]], i32 [[T31]], i64 7
437; AVX-NEXT:    [[TMP9:%.*]] = add <8 x i32> [[TMP8]], <i32 1, i32 2, i32 3, i32 4, i32 1, i32 2, i32 3, i32 4>
438; AVX-NEXT:    [[TMP10:%.*]] = bitcast i32* [[T0:%.*]] to <8 x i32>*
439; AVX-NEXT:    store <8 x i32> [[TMP9]], <8 x i32>* [[TMP10]], align 4, !tbaa [[TBAA0]]
440; AVX-NEXT:    ret void
441;
442; AVX2-LABEL: @gather_load_4(
443; AVX2-NEXT:    [[T6:%.*]] = getelementptr inbounds i32, i32* [[T1:%.*]], i64 11
444; AVX2-NEXT:    [[T10:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 4
445; AVX2-NEXT:    [[T14:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 15
446; AVX2-NEXT:    [[T18:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 18
447; AVX2-NEXT:    [[T22:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 9
448; AVX2-NEXT:    [[T26:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 6
449; AVX2-NEXT:    [[T30:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 21
450; AVX2-NEXT:    [[T3:%.*]] = load i32, i32* [[T1]], align 4, !tbaa [[TBAA0]]
451; AVX2-NEXT:    [[T7:%.*]] = load i32, i32* [[T6]], align 4, !tbaa [[TBAA0]]
452; AVX2-NEXT:    [[T11:%.*]] = load i32, i32* [[T10]], align 4, !tbaa [[TBAA0]]
453; AVX2-NEXT:    [[T15:%.*]] = load i32, i32* [[T14]], align 4, !tbaa [[TBAA0]]
454; AVX2-NEXT:    [[T19:%.*]] = load i32, i32* [[T18]], align 4, !tbaa [[TBAA0]]
455; AVX2-NEXT:    [[T23:%.*]] = load i32, i32* [[T22]], align 4, !tbaa [[TBAA0]]
456; AVX2-NEXT:    [[T27:%.*]] = load i32, i32* [[T26]], align 4, !tbaa [[TBAA0]]
457; AVX2-NEXT:    [[T31:%.*]] = load i32, i32* [[T30]], align 4, !tbaa [[TBAA0]]
458; AVX2-NEXT:    [[TMP1:%.*]] = insertelement <8 x i32> poison, i32 [[T3]], i64 0
459; AVX2-NEXT:    [[TMP2:%.*]] = insertelement <8 x i32> [[TMP1]], i32 [[T7]], i64 1
460; AVX2-NEXT:    [[TMP3:%.*]] = insertelement <8 x i32> [[TMP2]], i32 [[T11]], i64 2
461; AVX2-NEXT:    [[TMP4:%.*]] = insertelement <8 x i32> [[TMP3]], i32 [[T15]], i64 3
462; AVX2-NEXT:    [[TMP5:%.*]] = insertelement <8 x i32> [[TMP4]], i32 [[T19]], i64 4
463; AVX2-NEXT:    [[TMP6:%.*]] = insertelement <8 x i32> [[TMP5]], i32 [[T23]], i64 5
464; AVX2-NEXT:    [[TMP7:%.*]] = insertelement <8 x i32> [[TMP6]], i32 [[T27]], i64 6
465; AVX2-NEXT:    [[TMP8:%.*]] = insertelement <8 x i32> [[TMP7]], i32 [[T31]], i64 7
466; AVX2-NEXT:    [[TMP9:%.*]] = add <8 x i32> [[TMP8]], <i32 1, i32 2, i32 3, i32 4, i32 1, i32 2, i32 3, i32 4>
467; AVX2-NEXT:    [[TMP10:%.*]] = bitcast i32* [[T0:%.*]] to <8 x i32>*
468; AVX2-NEXT:    store <8 x i32> [[TMP9]], <8 x i32>* [[TMP10]], align 4, !tbaa [[TBAA0]]
469; AVX2-NEXT:    ret void
470;
471; AVX512F-LABEL: @gather_load_4(
472; AVX512F-NEXT:    [[T5:%.*]] = getelementptr inbounds i32, i32* [[T0:%.*]], i64 1
473; AVX512F-NEXT:    [[T6:%.*]] = getelementptr inbounds i32, i32* [[T1:%.*]], i64 11
474; AVX512F-NEXT:    [[T9:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 2
475; AVX512F-NEXT:    [[T10:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 4
476; AVX512F-NEXT:    [[T13:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 3
477; AVX512F-NEXT:    [[T14:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 15
478; AVX512F-NEXT:    [[T17:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 4
479; AVX512F-NEXT:    [[T18:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 18
480; AVX512F-NEXT:    [[T21:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 5
481; AVX512F-NEXT:    [[T22:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 9
482; AVX512F-NEXT:    [[T25:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 6
483; AVX512F-NEXT:    [[T26:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 6
484; AVX512F-NEXT:    [[T29:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 7
485; AVX512F-NEXT:    [[T30:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 21
486; AVX512F-NEXT:    [[T3:%.*]] = load i32, i32* [[T1]], align 4, !tbaa [[TBAA0]]
487; AVX512F-NEXT:    [[T7:%.*]] = load i32, i32* [[T6]], align 4, !tbaa [[TBAA0]]
488; AVX512F-NEXT:    [[T11:%.*]] = load i32, i32* [[T10]], align 4, !tbaa [[TBAA0]]
489; AVX512F-NEXT:    [[T15:%.*]] = load i32, i32* [[T14]], align 4, !tbaa [[TBAA0]]
490; AVX512F-NEXT:    [[T19:%.*]] = load i32, i32* [[T18]], align 4, !tbaa [[TBAA0]]
491; AVX512F-NEXT:    [[T23:%.*]] = load i32, i32* [[T22]], align 4, !tbaa [[TBAA0]]
492; AVX512F-NEXT:    [[T27:%.*]] = load i32, i32* [[T26]], align 4, !tbaa [[TBAA0]]
493; AVX512F-NEXT:    [[T31:%.*]] = load i32, i32* [[T30]], align 4, !tbaa [[TBAA0]]
494; AVX512F-NEXT:    [[T4:%.*]] = add i32 [[T3]], 1
495; AVX512F-NEXT:    [[T8:%.*]] = add i32 [[T7]], 2
496; AVX512F-NEXT:    [[T12:%.*]] = add i32 [[T11]], 3
497; AVX512F-NEXT:    [[T16:%.*]] = add i32 [[T15]], 4
498; AVX512F-NEXT:    [[T20:%.*]] = add i32 [[T19]], 1
499; AVX512F-NEXT:    [[T24:%.*]] = add i32 [[T23]], 2
500; AVX512F-NEXT:    [[T28:%.*]] = add i32 [[T27]], 3
501; AVX512F-NEXT:    [[T32:%.*]] = add i32 [[T31]], 4
502; AVX512F-NEXT:    store i32 [[T4]], i32* [[T0]], align 4, !tbaa [[TBAA0]]
503; AVX512F-NEXT:    store i32 [[T8]], i32* [[T5]], align 4, !tbaa [[TBAA0]]
504; AVX512F-NEXT:    store i32 [[T12]], i32* [[T9]], align 4, !tbaa [[TBAA0]]
505; AVX512F-NEXT:    store i32 [[T16]], i32* [[T13]], align 4, !tbaa [[TBAA0]]
506; AVX512F-NEXT:    store i32 [[T20]], i32* [[T17]], align 4, !tbaa [[TBAA0]]
507; AVX512F-NEXT:    store i32 [[T24]], i32* [[T21]], align 4, !tbaa [[TBAA0]]
508; AVX512F-NEXT:    store i32 [[T28]], i32* [[T25]], align 4, !tbaa [[TBAA0]]
509; AVX512F-NEXT:    store i32 [[T32]], i32* [[T29]], align 4, !tbaa [[TBAA0]]
510; AVX512F-NEXT:    ret void
511;
512; AVX512VL-LABEL: @gather_load_4(
513; AVX512VL-NEXT:    [[T5:%.*]] = getelementptr inbounds i32, i32* [[T0:%.*]], i64 1
514; AVX512VL-NEXT:    [[TMP1:%.*]] = insertelement <4 x i32*> poison, i32* [[T1:%.*]], i64 0
515; AVX512VL-NEXT:    [[SHUFFLE:%.*]] = shufflevector <4 x i32*> [[TMP1]], <4 x i32*> poison, <4 x i32> zeroinitializer
516; AVX512VL-NEXT:    [[TMP2:%.*]] = getelementptr i32, <4 x i32*> [[SHUFFLE]], <4 x i64> <i64 11, i64 4, i64 15, i64 18>
517; AVX512VL-NEXT:    [[T21:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 5
518; AVX512VL-NEXT:    [[T22:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 9
519; AVX512VL-NEXT:    [[T25:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 6
520; AVX512VL-NEXT:    [[T26:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 6
521; AVX512VL-NEXT:    [[T29:%.*]] = getelementptr inbounds i32, i32* [[T0]], i64 7
522; AVX512VL-NEXT:    [[T30:%.*]] = getelementptr inbounds i32, i32* [[T1]], i64 21
523; AVX512VL-NEXT:    [[T3:%.*]] = load i32, i32* [[T1]], align 4, !tbaa [[TBAA0]]
524; AVX512VL-NEXT:    [[TMP3:%.*]] = call <4 x i32> @llvm.masked.gather.v4i32.v4p0i32(<4 x i32*> [[TMP2]], i32 4, <4 x i1> <i1 true, i1 true, i1 true, i1 true>, <4 x i32> undef), !tbaa [[TBAA0]]
525; AVX512VL-NEXT:    [[T23:%.*]] = load i32, i32* [[T22]], align 4, !tbaa [[TBAA0]]
526; AVX512VL-NEXT:    [[T27:%.*]] = load i32, i32* [[T26]], align 4, !tbaa [[TBAA0]]
527; AVX512VL-NEXT:    [[T31:%.*]] = load i32, i32* [[T30]], align 4, !tbaa [[TBAA0]]
528; AVX512VL-NEXT:    [[T4:%.*]] = add i32 [[T3]], 1
529; AVX512VL-NEXT:    [[TMP4:%.*]] = add <4 x i32> [[TMP3]], <i32 2, i32 3, i32 4, i32 1>
530; AVX512VL-NEXT:    [[T24:%.*]] = add i32 [[T23]], 2
531; AVX512VL-NEXT:    [[T28:%.*]] = add i32 [[T27]], 3
532; AVX512VL-NEXT:    [[T32:%.*]] = add i32 [[T31]], 4
533; AVX512VL-NEXT:    store i32 [[T4]], i32* [[T0]], align 4, !tbaa [[TBAA0]]
534; AVX512VL-NEXT:    [[TMP5:%.*]] = bitcast i32* [[T5]] to <4 x i32>*
535; AVX512VL-NEXT:    store <4 x i32> [[TMP4]], <4 x i32>* [[TMP5]], align 4, !tbaa [[TBAA0]]
536; AVX512VL-NEXT:    store i32 [[T24]], i32* [[T21]], align 4, !tbaa [[TBAA0]]
537; AVX512VL-NEXT:    store i32 [[T28]], i32* [[T25]], align 4, !tbaa [[TBAA0]]
538; AVX512VL-NEXT:    store i32 [[T32]], i32* [[T29]], align 4, !tbaa [[TBAA0]]
539; AVX512VL-NEXT:    ret void
540;
541  %t5 = getelementptr inbounds i32, i32* %t0, i64 1
542  %t6 = getelementptr inbounds i32, i32* %t1, i64 11
543  %t9 = getelementptr inbounds i32, i32* %t0, i64 2
544  %t10 = getelementptr inbounds i32, i32* %t1, i64 4
545  %t13 = getelementptr inbounds i32, i32* %t0, i64 3
546  %t14 = getelementptr inbounds i32, i32* %t1, i64 15
547  %t17 = getelementptr inbounds i32, i32* %t0, i64 4
548  %t18 = getelementptr inbounds i32, i32* %t1, i64 18
549  %t21 = getelementptr inbounds i32, i32* %t0, i64 5
550  %t22 = getelementptr inbounds i32, i32* %t1, i64 9
551  %t25 = getelementptr inbounds i32, i32* %t0, i64 6
552  %t26 = getelementptr inbounds i32, i32* %t1, i64 6
553  %t29 = getelementptr inbounds i32, i32* %t0, i64 7
554  %t30 = getelementptr inbounds i32, i32* %t1, i64 21
555
556  %t3 = load i32, i32* %t1, align 4, !tbaa !2
557  %t7 = load i32, i32* %t6, align 4, !tbaa !2
558  %t11 = load i32, i32* %t10, align 4, !tbaa !2
559  %t15 = load i32, i32* %t14, align 4, !tbaa !2
560  %t19 = load i32, i32* %t18, align 4, !tbaa !2
561  %t23 = load i32, i32* %t22, align 4, !tbaa !2
562  %t27 = load i32, i32* %t26, align 4, !tbaa !2
563  %t31 = load i32, i32* %t30, align 4, !tbaa !2
564
565  %t4 = add i32 %t3, 1
566  %t8 = add i32 %t7, 2
567  %t12 = add i32 %t11, 3
568  %t16 = add i32 %t15, 4
569  %t20 = add i32 %t19, 1
570  %t24 = add i32 %t23, 2
571  %t28 = add i32 %t27, 3
572  %t32 = add i32 %t31, 4
573
574  store i32 %t4, i32* %t0, align 4, !tbaa !2
575  store i32 %t8, i32* %t5, align 4, !tbaa !2
576  store i32 %t12, i32* %t9, align 4, !tbaa !2
577  store i32 %t16, i32* %t13, align 4, !tbaa !2
578  store i32 %t20, i32* %t17, align 4, !tbaa !2
579  store i32 %t24, i32* %t21, align 4, !tbaa !2
580  store i32 %t28, i32* %t25, align 4, !tbaa !2
581  store i32 %t32, i32* %t29, align 4, !tbaa !2
582
583  ret void
584}
585
586
587define void @gather_load_div(float* noalias nocapture %0, float* noalias nocapture readonly %1) {
588; SSE-LABEL: @gather_load_div(
589; SSE-NEXT:    [[TMP3:%.*]] = load float, float* [[TMP1:%.*]], align 4, !tbaa [[TBAA0]]
590; SSE-NEXT:    [[TMP4:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 4
591; SSE-NEXT:    [[TMP5:%.*]] = load float, float* [[TMP4]], align 4, !tbaa [[TBAA0]]
592; SSE-NEXT:    [[TMP6:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 10
593; SSE-NEXT:    [[TMP7:%.*]] = load float, float* [[TMP6]], align 4, !tbaa [[TBAA0]]
594; SSE-NEXT:    [[TMP8:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 13
595; SSE-NEXT:    [[TMP9:%.*]] = load float, float* [[TMP8]], align 4, !tbaa [[TBAA0]]
596; SSE-NEXT:    [[TMP10:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 3
597; SSE-NEXT:    [[TMP11:%.*]] = load float, float* [[TMP10]], align 4, !tbaa [[TBAA0]]
598; SSE-NEXT:    [[TMP12:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 11
599; SSE-NEXT:    [[TMP13:%.*]] = load float, float* [[TMP12]], align 4, !tbaa [[TBAA0]]
600; SSE-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 14
601; SSE-NEXT:    [[TMP15:%.*]] = load float, float* [[TMP14]], align 4, !tbaa [[TBAA0]]
602; SSE-NEXT:    [[TMP16:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 44
603; SSE-NEXT:    [[TMP17:%.*]] = load float, float* [[TMP16]], align 4, !tbaa [[TBAA0]]
604; SSE-NEXT:    [[TMP18:%.*]] = insertelement <4 x float> poison, float [[TMP3]], i64 0
605; SSE-NEXT:    [[TMP19:%.*]] = insertelement <4 x float> [[TMP18]], float [[TMP7]], i64 1
606; SSE-NEXT:    [[TMP20:%.*]] = insertelement <4 x float> [[TMP19]], float [[TMP11]], i64 2
607; SSE-NEXT:    [[TMP21:%.*]] = insertelement <4 x float> [[TMP20]], float [[TMP15]], i64 3
608; SSE-NEXT:    [[TMP22:%.*]] = insertelement <4 x float> poison, float [[TMP5]], i64 0
609; SSE-NEXT:    [[TMP23:%.*]] = insertelement <4 x float> [[TMP22]], float [[TMP9]], i64 1
610; SSE-NEXT:    [[TMP24:%.*]] = insertelement <4 x float> [[TMP23]], float [[TMP13]], i64 2
611; SSE-NEXT:    [[TMP25:%.*]] = insertelement <4 x float> [[TMP24]], float [[TMP17]], i64 3
612; SSE-NEXT:    [[TMP26:%.*]] = fdiv <4 x float> [[TMP21]], [[TMP25]]
613; SSE-NEXT:    [[TMP27:%.*]] = getelementptr inbounds float, float* [[TMP0:%.*]], i64 4
614; SSE-NEXT:    [[TMP28:%.*]] = bitcast float* [[TMP0]] to <4 x float>*
615; SSE-NEXT:    store <4 x float> [[TMP26]], <4 x float>* [[TMP28]], align 4, !tbaa [[TBAA0]]
616; SSE-NEXT:    [[TMP29:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 17
617; SSE-NEXT:    [[TMP30:%.*]] = load float, float* [[TMP29]], align 4, !tbaa [[TBAA0]]
618; SSE-NEXT:    [[TMP31:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 33
619; SSE-NEXT:    [[TMP32:%.*]] = load float, float* [[TMP31]], align 4, !tbaa [[TBAA0]]
620; SSE-NEXT:    [[TMP33:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 8
621; SSE-NEXT:    [[TMP34:%.*]] = load float, float* [[TMP33]], align 4, !tbaa [[TBAA0]]
622; SSE-NEXT:    [[TMP35:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 30
623; SSE-NEXT:    [[TMP36:%.*]] = load float, float* [[TMP35]], align 4, !tbaa [[TBAA0]]
624; SSE-NEXT:    [[TMP37:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 5
625; SSE-NEXT:    [[TMP38:%.*]] = load float, float* [[TMP37]], align 4, !tbaa [[TBAA0]]
626; SSE-NEXT:    [[TMP39:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 27
627; SSE-NEXT:    [[TMP40:%.*]] = load float, float* [[TMP39]], align 4, !tbaa [[TBAA0]]
628; SSE-NEXT:    [[TMP41:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 20
629; SSE-NEXT:    [[TMP42:%.*]] = load float, float* [[TMP41]], align 4, !tbaa [[TBAA0]]
630; SSE-NEXT:    [[TMP43:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 23
631; SSE-NEXT:    [[TMP44:%.*]] = load float, float* [[TMP43]], align 4, !tbaa [[TBAA0]]
632; SSE-NEXT:    [[TMP45:%.*]] = insertelement <4 x float> poison, float [[TMP30]], i64 0
633; SSE-NEXT:    [[TMP46:%.*]] = insertelement <4 x float> [[TMP45]], float [[TMP34]], i64 1
634; SSE-NEXT:    [[TMP47:%.*]] = insertelement <4 x float> [[TMP46]], float [[TMP38]], i64 2
635; SSE-NEXT:    [[TMP48:%.*]] = insertelement <4 x float> [[TMP47]], float [[TMP42]], i64 3
636; SSE-NEXT:    [[TMP49:%.*]] = insertelement <4 x float> poison, float [[TMP32]], i64 0
637; SSE-NEXT:    [[TMP50:%.*]] = insertelement <4 x float> [[TMP49]], float [[TMP36]], i64 1
638; SSE-NEXT:    [[TMP51:%.*]] = insertelement <4 x float> [[TMP50]], float [[TMP40]], i64 2
639; SSE-NEXT:    [[TMP52:%.*]] = insertelement <4 x float> [[TMP51]], float [[TMP44]], i64 3
640; SSE-NEXT:    [[TMP53:%.*]] = fdiv <4 x float> [[TMP48]], [[TMP52]]
641; SSE-NEXT:    [[TMP54:%.*]] = bitcast float* [[TMP27]] to <4 x float>*
642; SSE-NEXT:    store <4 x float> [[TMP53]], <4 x float>* [[TMP54]], align 4, !tbaa [[TBAA0]]
643; SSE-NEXT:    ret void
644;
645; AVX-LABEL: @gather_load_div(
646; AVX-NEXT:    [[TMP3:%.*]] = load float, float* [[TMP1:%.*]], align 4, !tbaa [[TBAA0]]
647; AVX-NEXT:    [[TMP4:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 4
648; AVX-NEXT:    [[TMP5:%.*]] = load float, float* [[TMP4]], align 4, !tbaa [[TBAA0]]
649; AVX-NEXT:    [[TMP6:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 10
650; AVX-NEXT:    [[TMP7:%.*]] = load float, float* [[TMP6]], align 4, !tbaa [[TBAA0]]
651; AVX-NEXT:    [[TMP8:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 13
652; AVX-NEXT:    [[TMP9:%.*]] = load float, float* [[TMP8]], align 4, !tbaa [[TBAA0]]
653; AVX-NEXT:    [[TMP10:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 3
654; AVX-NEXT:    [[TMP11:%.*]] = load float, float* [[TMP10]], align 4, !tbaa [[TBAA0]]
655; AVX-NEXT:    [[TMP12:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 11
656; AVX-NEXT:    [[TMP13:%.*]] = load float, float* [[TMP12]], align 4, !tbaa [[TBAA0]]
657; AVX-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 14
658; AVX-NEXT:    [[TMP15:%.*]] = load float, float* [[TMP14]], align 4, !tbaa [[TBAA0]]
659; AVX-NEXT:    [[TMP16:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 44
660; AVX-NEXT:    [[TMP17:%.*]] = load float, float* [[TMP16]], align 4, !tbaa [[TBAA0]]
661; AVX-NEXT:    [[TMP18:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 17
662; AVX-NEXT:    [[TMP19:%.*]] = load float, float* [[TMP18]], align 4, !tbaa [[TBAA0]]
663; AVX-NEXT:    [[TMP20:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 33
664; AVX-NEXT:    [[TMP21:%.*]] = load float, float* [[TMP20]], align 4, !tbaa [[TBAA0]]
665; AVX-NEXT:    [[TMP22:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 8
666; AVX-NEXT:    [[TMP23:%.*]] = load float, float* [[TMP22]], align 4, !tbaa [[TBAA0]]
667; AVX-NEXT:    [[TMP24:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 30
668; AVX-NEXT:    [[TMP25:%.*]] = load float, float* [[TMP24]], align 4, !tbaa [[TBAA0]]
669; AVX-NEXT:    [[TMP26:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 5
670; AVX-NEXT:    [[TMP27:%.*]] = load float, float* [[TMP26]], align 4, !tbaa [[TBAA0]]
671; AVX-NEXT:    [[TMP28:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 27
672; AVX-NEXT:    [[TMP29:%.*]] = load float, float* [[TMP28]], align 4, !tbaa [[TBAA0]]
673; AVX-NEXT:    [[TMP30:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 20
674; AVX-NEXT:    [[TMP31:%.*]] = load float, float* [[TMP30]], align 4, !tbaa [[TBAA0]]
675; AVX-NEXT:    [[TMP32:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 23
676; AVX-NEXT:    [[TMP33:%.*]] = load float, float* [[TMP32]], align 4, !tbaa [[TBAA0]]
677; AVX-NEXT:    [[TMP34:%.*]] = insertelement <8 x float> poison, float [[TMP3]], i64 0
678; AVX-NEXT:    [[TMP35:%.*]] = insertelement <8 x float> [[TMP34]], float [[TMP7]], i64 1
679; AVX-NEXT:    [[TMP36:%.*]] = insertelement <8 x float> [[TMP35]], float [[TMP11]], i64 2
680; AVX-NEXT:    [[TMP37:%.*]] = insertelement <8 x float> [[TMP36]], float [[TMP15]], i64 3
681; AVX-NEXT:    [[TMP38:%.*]] = insertelement <8 x float> [[TMP37]], float [[TMP19]], i64 4
682; AVX-NEXT:    [[TMP39:%.*]] = insertelement <8 x float> [[TMP38]], float [[TMP23]], i64 5
683; AVX-NEXT:    [[TMP40:%.*]] = insertelement <8 x float> [[TMP39]], float [[TMP27]], i64 6
684; AVX-NEXT:    [[TMP41:%.*]] = insertelement <8 x float> [[TMP40]], float [[TMP31]], i64 7
685; AVX-NEXT:    [[TMP42:%.*]] = insertelement <8 x float> poison, float [[TMP5]], i64 0
686; AVX-NEXT:    [[TMP43:%.*]] = insertelement <8 x float> [[TMP42]], float [[TMP9]], i64 1
687; AVX-NEXT:    [[TMP44:%.*]] = insertelement <8 x float> [[TMP43]], float [[TMP13]], i64 2
688; AVX-NEXT:    [[TMP45:%.*]] = insertelement <8 x float> [[TMP44]], float [[TMP17]], i64 3
689; AVX-NEXT:    [[TMP46:%.*]] = insertelement <8 x float> [[TMP45]], float [[TMP21]], i64 4
690; AVX-NEXT:    [[TMP47:%.*]] = insertelement <8 x float> [[TMP46]], float [[TMP25]], i64 5
691; AVX-NEXT:    [[TMP48:%.*]] = insertelement <8 x float> [[TMP47]], float [[TMP29]], i64 6
692; AVX-NEXT:    [[TMP49:%.*]] = insertelement <8 x float> [[TMP48]], float [[TMP33]], i64 7
693; AVX-NEXT:    [[TMP50:%.*]] = fdiv <8 x float> [[TMP41]], [[TMP49]]
694; AVX-NEXT:    [[TMP51:%.*]] = bitcast float* [[TMP0:%.*]] to <8 x float>*
695; AVX-NEXT:    store <8 x float> [[TMP50]], <8 x float>* [[TMP51]], align 4, !tbaa [[TBAA0]]
696; AVX-NEXT:    ret void
697;
698; AVX2-LABEL: @gather_load_div(
699; AVX2-NEXT:    [[TMP3:%.*]] = load float, float* [[TMP1:%.*]], align 4, !tbaa [[TBAA0]]
700; AVX2-NEXT:    [[TMP4:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 4
701; AVX2-NEXT:    [[TMP5:%.*]] = load float, float* [[TMP4]], align 4, !tbaa [[TBAA0]]
702; AVX2-NEXT:    [[TMP6:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 10
703; AVX2-NEXT:    [[TMP7:%.*]] = load float, float* [[TMP6]], align 4, !tbaa [[TBAA0]]
704; AVX2-NEXT:    [[TMP8:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 13
705; AVX2-NEXT:    [[TMP9:%.*]] = load float, float* [[TMP8]], align 4, !tbaa [[TBAA0]]
706; AVX2-NEXT:    [[TMP10:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 3
707; AVX2-NEXT:    [[TMP11:%.*]] = load float, float* [[TMP10]], align 4, !tbaa [[TBAA0]]
708; AVX2-NEXT:    [[TMP12:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 11
709; AVX2-NEXT:    [[TMP13:%.*]] = load float, float* [[TMP12]], align 4, !tbaa [[TBAA0]]
710; AVX2-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 14
711; AVX2-NEXT:    [[TMP15:%.*]] = load float, float* [[TMP14]], align 4, !tbaa [[TBAA0]]
712; AVX2-NEXT:    [[TMP16:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 44
713; AVX2-NEXT:    [[TMP17:%.*]] = load float, float* [[TMP16]], align 4, !tbaa [[TBAA0]]
714; AVX2-NEXT:    [[TMP18:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 17
715; AVX2-NEXT:    [[TMP19:%.*]] = load float, float* [[TMP18]], align 4, !tbaa [[TBAA0]]
716; AVX2-NEXT:    [[TMP20:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 33
717; AVX2-NEXT:    [[TMP21:%.*]] = load float, float* [[TMP20]], align 4, !tbaa [[TBAA0]]
718; AVX2-NEXT:    [[TMP22:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 8
719; AVX2-NEXT:    [[TMP23:%.*]] = load float, float* [[TMP22]], align 4, !tbaa [[TBAA0]]
720; AVX2-NEXT:    [[TMP24:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 30
721; AVX2-NEXT:    [[TMP25:%.*]] = load float, float* [[TMP24]], align 4, !tbaa [[TBAA0]]
722; AVX2-NEXT:    [[TMP26:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 5
723; AVX2-NEXT:    [[TMP27:%.*]] = load float, float* [[TMP26]], align 4, !tbaa [[TBAA0]]
724; AVX2-NEXT:    [[TMP28:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 27
725; AVX2-NEXT:    [[TMP29:%.*]] = load float, float* [[TMP28]], align 4, !tbaa [[TBAA0]]
726; AVX2-NEXT:    [[TMP30:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 20
727; AVX2-NEXT:    [[TMP31:%.*]] = load float, float* [[TMP30]], align 4, !tbaa [[TBAA0]]
728; AVX2-NEXT:    [[TMP32:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 23
729; AVX2-NEXT:    [[TMP33:%.*]] = load float, float* [[TMP32]], align 4, !tbaa [[TBAA0]]
730; AVX2-NEXT:    [[TMP34:%.*]] = insertelement <8 x float> poison, float [[TMP3]], i64 0
731; AVX2-NEXT:    [[TMP35:%.*]] = insertelement <8 x float> [[TMP34]], float [[TMP7]], i64 1
732; AVX2-NEXT:    [[TMP36:%.*]] = insertelement <8 x float> [[TMP35]], float [[TMP11]], i64 2
733; AVX2-NEXT:    [[TMP37:%.*]] = insertelement <8 x float> [[TMP36]], float [[TMP15]], i64 3
734; AVX2-NEXT:    [[TMP38:%.*]] = insertelement <8 x float> [[TMP37]], float [[TMP19]], i64 4
735; AVX2-NEXT:    [[TMP39:%.*]] = insertelement <8 x float> [[TMP38]], float [[TMP23]], i64 5
736; AVX2-NEXT:    [[TMP40:%.*]] = insertelement <8 x float> [[TMP39]], float [[TMP27]], i64 6
737; AVX2-NEXT:    [[TMP41:%.*]] = insertelement <8 x float> [[TMP40]], float [[TMP31]], i64 7
738; AVX2-NEXT:    [[TMP42:%.*]] = insertelement <8 x float> poison, float [[TMP5]], i64 0
739; AVX2-NEXT:    [[TMP43:%.*]] = insertelement <8 x float> [[TMP42]], float [[TMP9]], i64 1
740; AVX2-NEXT:    [[TMP44:%.*]] = insertelement <8 x float> [[TMP43]], float [[TMP13]], i64 2
741; AVX2-NEXT:    [[TMP45:%.*]] = insertelement <8 x float> [[TMP44]], float [[TMP17]], i64 3
742; AVX2-NEXT:    [[TMP46:%.*]] = insertelement <8 x float> [[TMP45]], float [[TMP21]], i64 4
743; AVX2-NEXT:    [[TMP47:%.*]] = insertelement <8 x float> [[TMP46]], float [[TMP25]], i64 5
744; AVX2-NEXT:    [[TMP48:%.*]] = insertelement <8 x float> [[TMP47]], float [[TMP29]], i64 6
745; AVX2-NEXT:    [[TMP49:%.*]] = insertelement <8 x float> [[TMP48]], float [[TMP33]], i64 7
746; AVX2-NEXT:    [[TMP50:%.*]] = fdiv <8 x float> [[TMP41]], [[TMP49]]
747; AVX2-NEXT:    [[TMP51:%.*]] = bitcast float* [[TMP0:%.*]] to <8 x float>*
748; AVX2-NEXT:    store <8 x float> [[TMP50]], <8 x float>* [[TMP51]], align 4, !tbaa [[TBAA0]]
749; AVX2-NEXT:    ret void
750;
751; AVX512F-LABEL: @gather_load_div(
752; AVX512F-NEXT:    [[TMP3:%.*]] = insertelement <8 x float*> poison, float* [[TMP1:%.*]], i64 0
753; AVX512F-NEXT:    [[SHUFFLE:%.*]] = shufflevector <8 x float*> [[TMP3]], <8 x float*> poison, <8 x i32> zeroinitializer
754; AVX512F-NEXT:    [[TMP4:%.*]] = getelementptr float, <8 x float*> [[SHUFFLE]], <8 x i64> <i64 4, i64 13, i64 11, i64 44, i64 33, i64 30, i64 27, i64 23>
755; AVX512F-NEXT:    [[TMP5:%.*]] = insertelement <4 x float*> poison, float* [[TMP1]], i64 0
756; AVX512F-NEXT:    [[SHUFFLE1:%.*]] = shufflevector <4 x float*> [[TMP5]], <4 x float*> poison, <4 x i32> zeroinitializer
757; AVX512F-NEXT:    [[TMP6:%.*]] = getelementptr float, <4 x float*> [[SHUFFLE1]], <4 x i64> <i64 10, i64 3, i64 14, i64 17>
758; AVX512F-NEXT:    [[TMP7:%.*]] = insertelement <2 x float*> poison, float* [[TMP1]], i64 0
759; AVX512F-NEXT:    [[TMP8:%.*]] = shufflevector <2 x float*> [[TMP7]], <2 x float*> poison, <2 x i32> zeroinitializer
760; AVX512F-NEXT:    [[TMP9:%.*]] = getelementptr float, <2 x float*> [[TMP8]], <2 x i64> <i64 8, i64 5>
761; AVX512F-NEXT:    [[TMP10:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 20
762; AVX512F-NEXT:    [[TMP11:%.*]] = shufflevector <4 x float*> [[TMP6]], <4 x float*> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef>
763; AVX512F-NEXT:    [[TMP12:%.*]] = shufflevector <8 x float*> [[TMP3]], <8 x float*> [[TMP11]], <8 x i32> <i32 0, i32 8, i32 9, i32 10, i32 11, i32 undef, i32 undef, i32 undef>
764; AVX512F-NEXT:    [[TMP13:%.*]] = shufflevector <2 x float*> [[TMP9]], <2 x float*> poison, <8 x i32> <i32 0, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
765; AVX512F-NEXT:    [[TMP14:%.*]] = shufflevector <8 x float*> [[TMP12]], <8 x float*> [[TMP13]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 8, i32 9, i32 undef>
766; AVX512F-NEXT:    [[TMP15:%.*]] = insertelement <8 x float*> [[TMP14]], float* [[TMP10]], i64 7
767; AVX512F-NEXT:    [[TMP16:%.*]] = call <8 x float> @llvm.masked.gather.v8f32.v8p0f32(<8 x float*> [[TMP15]], i32 4, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x float> undef), !tbaa [[TBAA0]]
768; AVX512F-NEXT:    [[TMP17:%.*]] = call <8 x float> @llvm.masked.gather.v8f32.v8p0f32(<8 x float*> [[TMP4]], i32 4, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x float> undef), !tbaa [[TBAA0]]
769; AVX512F-NEXT:    [[TMP18:%.*]] = fdiv <8 x float> [[TMP16]], [[TMP17]]
770; AVX512F-NEXT:    [[TMP19:%.*]] = bitcast float* [[TMP0:%.*]] to <8 x float>*
771; AVX512F-NEXT:    store <8 x float> [[TMP18]], <8 x float>* [[TMP19]], align 4, !tbaa [[TBAA0]]
772; AVX512F-NEXT:    ret void
773;
774; AVX512VL-LABEL: @gather_load_div(
775; AVX512VL-NEXT:    [[TMP3:%.*]] = insertelement <8 x float*> poison, float* [[TMP1:%.*]], i64 0
776; AVX512VL-NEXT:    [[SHUFFLE:%.*]] = shufflevector <8 x float*> [[TMP3]], <8 x float*> poison, <8 x i32> zeroinitializer
777; AVX512VL-NEXT:    [[TMP4:%.*]] = getelementptr float, <8 x float*> [[SHUFFLE]], <8 x i64> <i64 4, i64 13, i64 11, i64 44, i64 33, i64 30, i64 27, i64 23>
778; AVX512VL-NEXT:    [[TMP5:%.*]] = insertelement <4 x float*> poison, float* [[TMP1]], i64 0
779; AVX512VL-NEXT:    [[SHUFFLE1:%.*]] = shufflevector <4 x float*> [[TMP5]], <4 x float*> poison, <4 x i32> zeroinitializer
780; AVX512VL-NEXT:    [[TMP6:%.*]] = getelementptr float, <4 x float*> [[SHUFFLE1]], <4 x i64> <i64 10, i64 3, i64 14, i64 17>
781; AVX512VL-NEXT:    [[TMP7:%.*]] = insertelement <2 x float*> poison, float* [[TMP1]], i64 0
782; AVX512VL-NEXT:    [[TMP8:%.*]] = shufflevector <2 x float*> [[TMP7]], <2 x float*> poison, <2 x i32> zeroinitializer
783; AVX512VL-NEXT:    [[TMP9:%.*]] = getelementptr float, <2 x float*> [[TMP8]], <2 x i64> <i64 8, i64 5>
784; AVX512VL-NEXT:    [[TMP10:%.*]] = getelementptr inbounds float, float* [[TMP1]], i64 20
785; AVX512VL-NEXT:    [[TMP11:%.*]] = shufflevector <4 x float*> [[TMP6]], <4 x float*> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef>
786; AVX512VL-NEXT:    [[TMP12:%.*]] = shufflevector <8 x float*> [[TMP3]], <8 x float*> [[TMP11]], <8 x i32> <i32 0, i32 8, i32 9, i32 10, i32 11, i32 undef, i32 undef, i32 undef>
787; AVX512VL-NEXT:    [[TMP13:%.*]] = shufflevector <2 x float*> [[TMP9]], <2 x float*> poison, <8 x i32> <i32 0, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
788; AVX512VL-NEXT:    [[TMP14:%.*]] = shufflevector <8 x float*> [[TMP12]], <8 x float*> [[TMP13]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 8, i32 9, i32 undef>
789; AVX512VL-NEXT:    [[TMP15:%.*]] = insertelement <8 x float*> [[TMP14]], float* [[TMP10]], i64 7
790; AVX512VL-NEXT:    [[TMP16:%.*]] = call <8 x float> @llvm.masked.gather.v8f32.v8p0f32(<8 x float*> [[TMP15]], i32 4, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x float> undef), !tbaa [[TBAA0]]
791; AVX512VL-NEXT:    [[TMP17:%.*]] = call <8 x float> @llvm.masked.gather.v8f32.v8p0f32(<8 x float*> [[TMP4]], i32 4, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x float> undef), !tbaa [[TBAA0]]
792; AVX512VL-NEXT:    [[TMP18:%.*]] = fdiv <8 x float> [[TMP16]], [[TMP17]]
793; AVX512VL-NEXT:    [[TMP19:%.*]] = bitcast float* [[TMP0:%.*]] to <8 x float>*
794; AVX512VL-NEXT:    store <8 x float> [[TMP18]], <8 x float>* [[TMP19]], align 4, !tbaa [[TBAA0]]
795; AVX512VL-NEXT:    ret void
796;
797; AVX512-LABEL: @gather_load_div(
798; AVX512-NEXT:    [[TMP3:%.*]] = getelementptr inbounds float, float* [[TMP1:%.*]], i64 10
799; AVX512-NEXT:    [[TMP4:%.*]] = insertelement <2 x float*> poison, float* [[TMP1]], i32 0
800; AVX512-NEXT:    [[TMP5:%.*]] = shufflevector <2 x float*> [[TMP4]], <2 x float*> poison, <2 x i32> zeroinitializer
801; AVX512-NEXT:    [[TMP6:%.*]] = getelementptr float, <2 x float*> [[TMP5]], <2 x i64> <i64 3, i64 14>
802; AVX512-NEXT:    [[TMP7:%.*]] = insertelement <4 x float*> poison, float* [[TMP1]], i32 0
803; AVX512-NEXT:    [[TMP8:%.*]] = shufflevector <4 x float*> [[TMP7]], <4 x float*> poison, <4 x i32> zeroinitializer
804; AVX512-NEXT:    [[TMP9:%.*]] = getelementptr float, <4 x float*> [[TMP8]], <4 x i64> <i64 17, i64 8, i64 5, i64 20>
805; AVX512-NEXT:    [[TMP10:%.*]] = insertelement <8 x float*> poison, float* [[TMP1]], i32 0
806; AVX512-NEXT:    [[TMP11:%.*]] = insertelement <8 x float*> [[TMP10]], float* [[TMP3]], i32 1
807; AVX512-NEXT:    [[TMP12:%.*]] = shufflevector <2 x float*> [[TMP6]], <2 x float*> poison, <8 x i32> <i32 0, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
808; AVX512-NEXT:    [[TMP13:%.*]] = shufflevector <8 x float*> [[TMP11]], <8 x float*> [[TMP12]], <8 x i32> <i32 0, i32 1, i32 8, i32 9, i32 undef, i32 undef, i32 undef, i32 undef>
809; AVX512-NEXT:    [[TMP14:%.*]] = shufflevector <4 x float*> [[TMP9]], <4 x float*> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef>
810; AVX512-NEXT:    [[TMP15:%.*]] = shufflevector <8 x float*> [[TMP13]], <8 x float*> [[TMP14]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
811; AVX512-NEXT:    [[TMP16:%.*]] = call <8 x float> @llvm.masked.gather.v8f32.v8p0f32(<8 x float*> [[TMP15]], i32 4, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x float> undef), !tbaa [[TBAA0]]
812; AVX512-NEXT:    [[TMP17:%.*]] = shufflevector <8 x float*> [[TMP10]], <8 x float*> poison, <8 x i32> zeroinitializer
813; AVX512-NEXT:    [[TMP18:%.*]] = getelementptr float, <8 x float*> [[TMP17]], <8 x i64> <i64 4, i64 13, i64 11, i64 44, i64 33, i64 30, i64 27, i64 23>
814; AVX512-NEXT:    [[TMP19:%.*]] = call <8 x float> @llvm.masked.gather.v8f32.v8p0f32(<8 x float*> [[TMP18]], i32 4, <8 x i1> <i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true, i1 true>, <8 x float> undef), !tbaa [[TBAA0]]
815; AVX512-NEXT:    [[TMP20:%.*]] = fdiv <8 x float> [[TMP16]], [[TMP19]]
816; AVX512-NEXT:    [[TMP21:%.*]] = bitcast float* [[TMP0:%.*]] to <8 x float>*
817; AVX512-NEXT:    store <8 x float> [[TMP20]], <8 x float>* [[TMP21]], align 4, !tbaa [[TBAA0]]
818; AVX512-NEXT:    ret void
819  %3 = load float, float* %1, align 4, !tbaa !2
820  %4 = getelementptr inbounds float, float* %1, i64 4
821  %5 = load float, float* %4, align 4, !tbaa !2
822  %6 = fdiv float %3, %5
823  %7 = getelementptr inbounds float, float* %0, i64 1
824  store float %6, float* %0, align 4, !tbaa !2
825  %8 = getelementptr inbounds float, float* %1, i64 10
826  %9 = load float, float* %8, align 4, !tbaa !2
827  %10 = getelementptr inbounds float, float* %1, i64 13
828  %11 = load float, float* %10, align 4, !tbaa !2
829  %12 = fdiv float %9, %11
830  %13 = getelementptr inbounds float, float* %0, i64 2
831  store float %12, float* %7, align 4, !tbaa !2
832  %14 = getelementptr inbounds float, float* %1, i64 3
833  %15 = load float, float* %14, align 4, !tbaa !2
834  %16 = getelementptr inbounds float, float* %1, i64 11
835  %17 = load float, float* %16, align 4, !tbaa !2
836  %18 = fdiv float %15, %17
837  %19 = getelementptr inbounds float, float* %0, i64 3
838  store float %18, float* %13, align 4, !tbaa !2
839  %20 = getelementptr inbounds float, float* %1, i64 14
840  %21 = load float, float* %20, align 4, !tbaa !2
841  %22 = getelementptr inbounds float, float* %1, i64 44
842  %23 = load float, float* %22, align 4, !tbaa !2
843  %24 = fdiv float %21, %23
844  %25 = getelementptr inbounds float, float* %0, i64 4
845  store float %24, float* %19, align 4, !tbaa !2
846  %26 = getelementptr inbounds float, float* %1, i64 17
847  %27 = load float, float* %26, align 4, !tbaa !2
848  %28 = getelementptr inbounds float, float* %1, i64 33
849  %29 = load float, float* %28, align 4, !tbaa !2
850  %30 = fdiv float %27, %29
851  %31 = getelementptr inbounds float, float* %0, i64 5
852  store float %30, float* %25, align 4, !tbaa !2
853  %32 = getelementptr inbounds float, float* %1, i64 8
854  %33 = load float, float* %32, align 4, !tbaa !2
855  %34 = getelementptr inbounds float, float* %1, i64 30
856  %35 = load float, float* %34, align 4, !tbaa !2
857  %36 = fdiv float %33, %35
858  %37 = getelementptr inbounds float, float* %0, i64 6
859  store float %36, float* %31, align 4, !tbaa !2
860  %38 = getelementptr inbounds float, float* %1, i64 5
861  %39 = load float, float* %38, align 4, !tbaa !2
862  %40 = getelementptr inbounds float, float* %1, i64 27
863  %41 = load float, float* %40, align 4, !tbaa !2
864  %42 = fdiv float %39, %41
865  %43 = getelementptr inbounds float, float* %0, i64 7
866  store float %42, float* %37, align 4, !tbaa !2
867  %44 = getelementptr inbounds float, float* %1, i64 20
868  %45 = load float, float* %44, align 4, !tbaa !2
869  %46 = getelementptr inbounds float, float* %1, i64 23
870  %47 = load float, float* %46, align 4, !tbaa !2
871  %48 = fdiv float %45, %47
872  store float %48, float* %43, align 4, !tbaa !2
873  ret void
874}
875
876!2 = !{!3, !3, i64 0}
877!3 = !{!"short", !4, i64 0}
878!4 = !{!"omnipotent char", !5, i64 0}
879!5 = !{!"Simple C++ TBAA"}
880