xref: /llvm-project/llvm/test/Transforms/LoopVectorize/PowerPC/optimal-epilog-vectorization.ll (revision 9cf67f6ea0ff3349ad4a1c094fd30434b4bb0475) 
1; RUN: opt < %s -passes='loop-vectorize' -enable-epilogue-vectorization -epilogue-vectorization-force-VF=2 -S | FileCheck %s --check-prefix VF-TWO-CHECK
2; RUN: opt < %s -passes='loop-vectorize' -enable-epilogue-vectorization -epilogue-vectorization-force-VF=4 -S | FileCheck %s --check-prefix VF-FOUR-CHECK
3
4target datalayout = "e-m:e-i64:64-n32:64"
5target triple = "powerpc64le-unknown-linux-gnu"
6
7; Function Attrs: nounwind
8define dso_local void @f1(ptr noalias %aa, ptr noalias %bb, ptr noalias %cc, i32 signext %N) #0 {
9; VF-TWO-CHECK-LABEL: @f1(
10; VF-TWO-CHECK-NEXT:  entry:
11; VF-TWO-CHECK-NEXT:    [[CMP1:%.*]] = icmp sgt i32 [[N:%.*]], 0
12; VF-TWO-CHECK-NEXT:    br i1 [[CMP1]], label [[ITER_CHECK:%.*]], label [[FOR_END:%.*]]
13; VF-TWO-CHECK:       iter.check:
14; VF-TWO-CHECK-NEXT:    [[WIDE_TRIP_COUNT:%.*]] = zext i32 [[N]] to i64
15; VF-TWO-CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[WIDE_TRIP_COUNT]], 2
16; VF-TWO-CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
17; VF-TWO-CHECK:       vector.main.loop.iter.check:
18; VF-TWO-CHECK-NEXT:    [[MIN_ITERS_CHECK1:%.*]] = icmp ult i64 [[WIDE_TRIP_COUNT]], 48
19; VF-TWO-CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
20; VF-TWO-CHECK:       vector.ph:
21; VF-TWO-CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[WIDE_TRIP_COUNT]], 48
22; VF-TWO-CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_MOD_VF]]
23; VF-TWO-CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
24; VF-TWO-CHECK:       vector.body:
25; VF-TWO-CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
26; VF-TWO-CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
27; VF-TWO-CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
28; VF-TWO-CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
29; VF-TWO-CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
30; VF-TWO-CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 16
31; VF-TWO-CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 20
32; VF-TWO-CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 24
33; VF-TWO-CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 28
34; VF-TWO-CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 32
35; VF-TWO-CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 36
36; VF-TWO-CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 40
37; VF-TWO-CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 44
38; VF-TWO-CHECK-NEXT:    [[TMP12:%.*]] = getelementptr inbounds float, ptr [[BB:%.*]], i64 [[TMP0]]
39; VF-TWO-CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP1]]
40; VF-TWO-CHECK-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP2]]
41; VF-TWO-CHECK-NEXT:    [[TMP15:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP3]]
42; VF-TWO-CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP4]]
43; VF-TWO-CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP5]]
44; VF-TWO-CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP6]]
45; VF-TWO-CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP7]]
46; VF-TWO-CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP8]]
47; VF-TWO-CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP9]]
48; VF-TWO-CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP10]]
49; VF-TWO-CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP11]]
50; VF-TWO-CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 0
51; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP24]], align 4
52; VF-TWO-CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 4
53; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD2:%.*]] = load <4 x float>, ptr [[TMP26]], align 4
54; VF-TWO-CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 8
55; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD3:%.*]] = load <4 x float>, ptr [[TMP28]], align 4
56; VF-TWO-CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 12
57; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x float>, ptr [[TMP30]], align 4
58; VF-TWO-CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 16
59; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x float>, ptr [[TMP32]], align 4
60; VF-TWO-CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 20
61; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x float>, ptr [[TMP34]], align 4
62; VF-TWO-CHECK-NEXT:    [[TMP36:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 24
63; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD7:%.*]] = load <4 x float>, ptr [[TMP36]], align 4
64; VF-TWO-CHECK-NEXT:    [[TMP38:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 28
65; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD8:%.*]] = load <4 x float>, ptr [[TMP38]], align 4
66; VF-TWO-CHECK-NEXT:    [[TMP40:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 32
67; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD9:%.*]] = load <4 x float>, ptr [[TMP40]], align 4
68; VF-TWO-CHECK-NEXT:    [[TMP42:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 36
69; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD10:%.*]] = load <4 x float>, ptr [[TMP42]], align 4
70; VF-TWO-CHECK-NEXT:    [[TMP44:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 40
71; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD11:%.*]] = load <4 x float>, ptr [[TMP44]], align 4
72; VF-TWO-CHECK-NEXT:    [[TMP46:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 44
73; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD12:%.*]] = load <4 x float>, ptr [[TMP46]], align 4
74; VF-TWO-CHECK-NEXT:    [[TMP48:%.*]] = getelementptr inbounds float, ptr [[CC:%.*]], i64 [[TMP0]]
75; VF-TWO-CHECK-NEXT:    [[TMP49:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP1]]
76; VF-TWO-CHECK-NEXT:    [[TMP50:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP2]]
77; VF-TWO-CHECK-NEXT:    [[TMP51:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP3]]
78; VF-TWO-CHECK-NEXT:    [[TMP52:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP4]]
79; VF-TWO-CHECK-NEXT:    [[TMP53:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP5]]
80; VF-TWO-CHECK-NEXT:    [[TMP54:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP6]]
81; VF-TWO-CHECK-NEXT:    [[TMP55:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP7]]
82; VF-TWO-CHECK-NEXT:    [[TMP56:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP8]]
83; VF-TWO-CHECK-NEXT:    [[TMP57:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP9]]
84; VF-TWO-CHECK-NEXT:    [[TMP58:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP10]]
85; VF-TWO-CHECK-NEXT:    [[TMP59:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP11]]
86; VF-TWO-CHECK-NEXT:    [[TMP60:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 0
87; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD13:%.*]] = load <4 x float>, ptr [[TMP60]], align 4
88; VF-TWO-CHECK-NEXT:    [[TMP62:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 4
89; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD14:%.*]] = load <4 x float>, ptr [[TMP62]], align 4
90; VF-TWO-CHECK-NEXT:    [[TMP64:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 8
91; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD15:%.*]] = load <4 x float>, ptr [[TMP64]], align 4
92; VF-TWO-CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 12
93; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD16:%.*]] = load <4 x float>, ptr [[TMP66]], align 4
94; VF-TWO-CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 16
95; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD17:%.*]] = load <4 x float>, ptr [[TMP68]], align 4
96; VF-TWO-CHECK-NEXT:    [[TMP70:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 20
97; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD18:%.*]] = load <4 x float>, ptr [[TMP70]], align 4
98; VF-TWO-CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 24
99; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD19:%.*]] = load <4 x float>, ptr [[TMP72]], align 4
100; VF-TWO-CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 28
101; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD20:%.*]] = load <4 x float>, ptr [[TMP74]], align 4
102; VF-TWO-CHECK-NEXT:    [[TMP76:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 32
103; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD21:%.*]] = load <4 x float>, ptr [[TMP76]], align 4
104; VF-TWO-CHECK-NEXT:    [[TMP78:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 36
105; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD22:%.*]] = load <4 x float>, ptr [[TMP78]], align 4
106; VF-TWO-CHECK-NEXT:    [[TMP80:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 40
107; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD23:%.*]] = load <4 x float>, ptr [[TMP80]], align 4
108; VF-TWO-CHECK-NEXT:    [[TMP82:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 44
109; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD24:%.*]] = load <4 x float>, ptr [[TMP82]], align 4
110; VF-TWO-CHECK-NEXT:    [[TMP84:%.*]] = fadd fast <4 x float> [[WIDE_LOAD]], [[WIDE_LOAD13]]
111; VF-TWO-CHECK-NEXT:    [[TMP85:%.*]] = fadd fast <4 x float> [[WIDE_LOAD2]], [[WIDE_LOAD14]]
112; VF-TWO-CHECK-NEXT:    [[TMP86:%.*]] = fadd fast <4 x float> [[WIDE_LOAD3]], [[WIDE_LOAD15]]
113; VF-TWO-CHECK-NEXT:    [[TMP87:%.*]] = fadd fast <4 x float> [[WIDE_LOAD4]], [[WIDE_LOAD16]]
114; VF-TWO-CHECK-NEXT:    [[TMP88:%.*]] = fadd fast <4 x float> [[WIDE_LOAD5]], [[WIDE_LOAD17]]
115; VF-TWO-CHECK-NEXT:    [[TMP89:%.*]] = fadd fast <4 x float> [[WIDE_LOAD6]], [[WIDE_LOAD18]]
116; VF-TWO-CHECK-NEXT:    [[TMP90:%.*]] = fadd fast <4 x float> [[WIDE_LOAD7]], [[WIDE_LOAD19]]
117; VF-TWO-CHECK-NEXT:    [[TMP91:%.*]] = fadd fast <4 x float> [[WIDE_LOAD8]], [[WIDE_LOAD20]]
118; VF-TWO-CHECK-NEXT:    [[TMP92:%.*]] = fadd fast <4 x float> [[WIDE_LOAD9]], [[WIDE_LOAD21]]
119; VF-TWO-CHECK-NEXT:    [[TMP93:%.*]] = fadd fast <4 x float> [[WIDE_LOAD10]], [[WIDE_LOAD22]]
120; VF-TWO-CHECK-NEXT:    [[TMP94:%.*]] = fadd fast <4 x float> [[WIDE_LOAD11]], [[WIDE_LOAD23]]
121; VF-TWO-CHECK-NEXT:    [[TMP95:%.*]] = fadd fast <4 x float> [[WIDE_LOAD12]], [[WIDE_LOAD24]]
122; VF-TWO-CHECK-NEXT:    [[TMP96:%.*]] = getelementptr inbounds float, ptr [[AA:%.*]], i64 [[TMP0]]
123; VF-TWO-CHECK-NEXT:    [[TMP97:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP1]]
124; VF-TWO-CHECK-NEXT:    [[TMP98:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP2]]
125; VF-TWO-CHECK-NEXT:    [[TMP99:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP3]]
126; VF-TWO-CHECK-NEXT:    [[TMP100:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP4]]
127; VF-TWO-CHECK-NEXT:    [[TMP101:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP5]]
128; VF-TWO-CHECK-NEXT:    [[TMP102:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP6]]
129; VF-TWO-CHECK-NEXT:    [[TMP103:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP7]]
130; VF-TWO-CHECK-NEXT:    [[TMP104:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP8]]
131; VF-TWO-CHECK-NEXT:    [[TMP105:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP9]]
132; VF-TWO-CHECK-NEXT:    [[TMP106:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP10]]
133; VF-TWO-CHECK-NEXT:    [[TMP107:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP11]]
134; VF-TWO-CHECK-NEXT:    [[TMP108:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 0
135; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP84]], ptr [[TMP108]], align 4
136; VF-TWO-CHECK-NEXT:    [[TMP110:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 4
137; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP85]], ptr [[TMP110]], align 4
138; VF-TWO-CHECK-NEXT:    [[TMP112:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 8
139; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP86]], ptr [[TMP112]], align 4
140; VF-TWO-CHECK-NEXT:    [[TMP114:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 12
141; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP87]], ptr [[TMP114]], align 4
142; VF-TWO-CHECK-NEXT:    [[TMP116:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 16
143; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP88]], ptr [[TMP116]], align 4
144; VF-TWO-CHECK-NEXT:    [[TMP118:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 20
145; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP89]], ptr [[TMP118]], align 4
146; VF-TWO-CHECK-NEXT:    [[TMP120:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 24
147; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP90]], ptr [[TMP120]], align 4
148; VF-TWO-CHECK-NEXT:    [[TMP122:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 28
149; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP91]], ptr [[TMP122]], align 4
150; VF-TWO-CHECK-NEXT:    [[TMP124:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 32
151; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP92]], ptr [[TMP124]], align 4
152; VF-TWO-CHECK-NEXT:    [[TMP126:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 36
153; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP93]], ptr [[TMP126]], align 4
154; VF-TWO-CHECK-NEXT:    [[TMP128:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 40
155; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP94]], ptr [[TMP128]], align 4
156; VF-TWO-CHECK-NEXT:    [[TMP130:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 44
157; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP95]], ptr [[TMP130]], align 4
158; VF-TWO-CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 48
159; VF-TWO-CHECK-NEXT:    [[TMP132:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
160; VF-TWO-CHECK-NEXT:    br i1 [[TMP132]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOPID_MV:![0-9]+]]
161; VF-TWO-CHECK:       middle.block:
162; VF-TWO-CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[WIDE_TRIP_COUNT]], [[N_VEC]]
163; VF-TWO-CHECK-NEXT:    br i1 [[CMP_N]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
164; VF-TWO-CHECK:       vec.epilog.iter.check:
165; VF-TWO-CHECK-NEXT:    [[N_VEC_REMAINING:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_VEC]]
166; VF-TWO-CHECK-NEXT:    [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_VEC_REMAINING]], 2
167; VF-TWO-CHECK-NEXT:    br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
168; VF-TWO-CHECK:       vec.epilog.ph:
169; VF-TWO-CHECK-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
170; VF-TWO-CHECK-NEXT:    [[N_MOD_VF25:%.*]] = urem i64 [[WIDE_TRIP_COUNT]], 2
171; VF-TWO-CHECK-NEXT:    [[N_VEC26:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_MOD_VF25]]
172; VF-TWO-CHECK-NEXT:    br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
173; VF-TWO-CHECK:       vec.epilog.vector.body:
174; VF-TWO-CHECK-NEXT:    [[OFFSET_IDX:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT31:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
175; VF-TWO-CHECK-NEXT:    [[TMP133:%.*]] = add i64 [[OFFSET_IDX]], 0
176; VF-TWO-CHECK-NEXT:    [[TMP134:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP133]]
177; VF-TWO-CHECK-NEXT:    [[TMP135:%.*]] = getelementptr inbounds float, ptr [[TMP134]], i32 0
178; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD29:%.*]] = load <2 x float>, ptr [[TMP135]], align 4
179; VF-TWO-CHECK-NEXT:    [[TMP137:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP133]]
180; VF-TWO-CHECK-NEXT:    [[TMP138:%.*]] = getelementptr inbounds float, ptr [[TMP137]], i32 0
181; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD30:%.*]] = load <2 x float>, ptr [[TMP138]], align 4
182; VF-TWO-CHECK-NEXT:    [[TMP140:%.*]] = fadd fast <2 x float> [[WIDE_LOAD29]], [[WIDE_LOAD30]]
183; VF-TWO-CHECK-NEXT:    [[TMP141:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP133]]
184; VF-TWO-CHECK-NEXT:    [[TMP142:%.*]] = getelementptr inbounds float, ptr [[TMP141]], i32 0
185; VF-TWO-CHECK-NEXT:    store <2 x float> [[TMP140]], ptr [[TMP142]], align 4
186; VF-TWO-CHECK-NEXT:    [[INDEX_NEXT31]] = add nuw i64 [[OFFSET_IDX]], 2
187; VF-TWO-CHECK-NEXT:    [[TMP144:%.*]] = icmp eq i64 [[INDEX_NEXT31]], [[N_VEC26]]
188; VF-TWO-CHECK-NEXT:    br i1 [[TMP144]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOPID_EV:![0-9]+]]
189; VF-TWO-CHECK:       vec.epilog.middle.block:
190; VF-TWO-CHECK-NEXT:    [[CMP_N27:%.*]] = icmp eq i64 [[WIDE_TRIP_COUNT]], [[N_VEC26]]
191; VF-TWO-CHECK-NEXT:    br i1 [[CMP_N27]], label [[FOR_END_LOOPEXIT]], label [[VEC_EPILOG_SCALAR_PH]]
192; VF-TWO-CHECK:       vec.epilog.scalar.ph:
193; VF-TWO-CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC26]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[ITER_CHECK]] ]
194; VF-TWO-CHECK-NEXT:    br label [[FOR_BODY:%.*]]
195; VF-TWO-CHECK:       for.body:
196; VF-TWO-CHECK-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
197; VF-TWO-CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[INDVARS_IV]]
198; VF-TWO-CHECK-NEXT:    [[TMP145:%.*]] = load float, ptr [[ARRAYIDX]], align 4
199; VF-TWO-CHECK-NEXT:    [[ARRAYIDX2:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[INDVARS_IV]]
200; VF-TWO-CHECK-NEXT:    [[TMP146:%.*]] = load float, ptr [[ARRAYIDX2]], align 4
201; VF-TWO-CHECK-NEXT:    [[ADD:%.*]] = fadd fast float [[TMP145]], [[TMP146]]
202; VF-TWO-CHECK-NEXT:    [[ARRAYIDX4:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[INDVARS_IV]]
203; VF-TWO-CHECK-NEXT:    store float [[ADD]], ptr [[ARRAYIDX4]], align 4
204; VF-TWO-CHECK-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
205; VF-TWO-CHECK-NEXT:    [[EXITCOND:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT]], [[WIDE_TRIP_COUNT]]
206; VF-TWO-CHECK-NEXT:    br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT]], !llvm.loop [[LOOP4:![0-9]+]]
207; VF-TWO-CHECK:       for.end.loopexit:
208; VF-TWO-CHECK-NEXT:    br label [[FOR_END]]
209; VF-TWO-CHECK:       for.end:
210; VF-TWO-CHECK-NEXT:    ret void
211;
212; VF-FOUR-CHECK-LABEL: @f1(
213; VF-FOUR-CHECK-NEXT:  entry:
214; VF-FOUR-CHECK-NEXT:    [[CMP1:%.*]] = icmp sgt i32 [[N:%.*]], 0
215; VF-FOUR-CHECK-NEXT:    br i1 [[CMP1]], label [[ITER_CHECK:%.*]], label [[FOR_END:%.*]]
216; VF-FOUR-CHECK:       iter.check:
217; VF-FOUR-CHECK-NEXT:    [[WIDE_TRIP_COUNT:%.*]] = zext i32 [[N]] to i64
218; VF-FOUR-CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[WIDE_TRIP_COUNT]], 4
219; VF-FOUR-CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
220; VF-FOUR-CHECK:       vector.main.loop.iter.check:
221; VF-FOUR-CHECK-NEXT:    [[MIN_ITERS_CHECK1:%.*]] = icmp ult i64 [[WIDE_TRIP_COUNT]], 48
222; VF-FOUR-CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
223; VF-FOUR-CHECK:       vector.ph:
224; VF-FOUR-CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[WIDE_TRIP_COUNT]], 48
225; VF-FOUR-CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_MOD_VF]]
226; VF-FOUR-CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
227; VF-FOUR-CHECK:       vector.body:
228; VF-FOUR-CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
229; VF-FOUR-CHECK-NEXT:    [[TMP0:%.*]] = add i64 [[INDEX]], 0
230; VF-FOUR-CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[INDEX]], 4
231; VF-FOUR-CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 8
232; VF-FOUR-CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 12
233; VF-FOUR-CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 16
234; VF-FOUR-CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 20
235; VF-FOUR-CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[INDEX]], 24
236; VF-FOUR-CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 28
237; VF-FOUR-CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[INDEX]], 32
238; VF-FOUR-CHECK-NEXT:    [[TMP9:%.*]] = add i64 [[INDEX]], 36
239; VF-FOUR-CHECK-NEXT:    [[TMP10:%.*]] = add i64 [[INDEX]], 40
240; VF-FOUR-CHECK-NEXT:    [[TMP11:%.*]] = add i64 [[INDEX]], 44
241; VF-FOUR-CHECK-NEXT:    [[TMP12:%.*]] = getelementptr inbounds float, ptr [[BB:%.*]], i64 [[TMP0]]
242; VF-FOUR-CHECK-NEXT:    [[TMP13:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP1]]
243; VF-FOUR-CHECK-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP2]]
244; VF-FOUR-CHECK-NEXT:    [[TMP15:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP3]]
245; VF-FOUR-CHECK-NEXT:    [[TMP16:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP4]]
246; VF-FOUR-CHECK-NEXT:    [[TMP17:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP5]]
247; VF-FOUR-CHECK-NEXT:    [[TMP18:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP6]]
248; VF-FOUR-CHECK-NEXT:    [[TMP19:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP7]]
249; VF-FOUR-CHECK-NEXT:    [[TMP20:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP8]]
250; VF-FOUR-CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP9]]
251; VF-FOUR-CHECK-NEXT:    [[TMP22:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP10]]
252; VF-FOUR-CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP11]]
253; VF-FOUR-CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 0
254; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP24]], align 4
255; VF-FOUR-CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 4
256; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD2:%.*]] = load <4 x float>, ptr [[TMP26]], align 4
257; VF-FOUR-CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 8
258; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD3:%.*]] = load <4 x float>, ptr [[TMP28]], align 4
259; VF-FOUR-CHECK-NEXT:    [[TMP30:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 12
260; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x float>, ptr [[TMP30]], align 4
261; VF-FOUR-CHECK-NEXT:    [[TMP32:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 16
262; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x float>, ptr [[TMP32]], align 4
263; VF-FOUR-CHECK-NEXT:    [[TMP34:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 20
264; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x float>, ptr [[TMP34]], align 4
265; VF-FOUR-CHECK-NEXT:    [[TMP36:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 24
266; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD7:%.*]] = load <4 x float>, ptr [[TMP36]], align 4
267; VF-FOUR-CHECK-NEXT:    [[TMP38:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 28
268; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD8:%.*]] = load <4 x float>, ptr [[TMP38]], align 4
269; VF-FOUR-CHECK-NEXT:    [[TMP40:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 32
270; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD9:%.*]] = load <4 x float>, ptr [[TMP40]], align 4
271; VF-FOUR-CHECK-NEXT:    [[TMP42:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 36
272; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD10:%.*]] = load <4 x float>, ptr [[TMP42]], align 4
273; VF-FOUR-CHECK-NEXT:    [[TMP44:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 40
274; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD11:%.*]] = load <4 x float>, ptr [[TMP44]], align 4
275; VF-FOUR-CHECK-NEXT:    [[TMP46:%.*]] = getelementptr inbounds float, ptr [[TMP12]], i32 44
276; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD12:%.*]] = load <4 x float>, ptr [[TMP46]], align 4
277; VF-FOUR-CHECK-NEXT:    [[TMP48:%.*]] = getelementptr inbounds float, ptr [[CC:%.*]], i64 [[TMP0]]
278; VF-FOUR-CHECK-NEXT:    [[TMP49:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP1]]
279; VF-FOUR-CHECK-NEXT:    [[TMP50:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP2]]
280; VF-FOUR-CHECK-NEXT:    [[TMP51:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP3]]
281; VF-FOUR-CHECK-NEXT:    [[TMP52:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP4]]
282; VF-FOUR-CHECK-NEXT:    [[TMP53:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP5]]
283; VF-FOUR-CHECK-NEXT:    [[TMP54:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP6]]
284; VF-FOUR-CHECK-NEXT:    [[TMP55:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP7]]
285; VF-FOUR-CHECK-NEXT:    [[TMP56:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP8]]
286; VF-FOUR-CHECK-NEXT:    [[TMP57:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP9]]
287; VF-FOUR-CHECK-NEXT:    [[TMP58:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP10]]
288; VF-FOUR-CHECK-NEXT:    [[TMP59:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP11]]
289; VF-FOUR-CHECK-NEXT:    [[TMP60:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 0
290; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD13:%.*]] = load <4 x float>, ptr [[TMP60]], align 4
291; VF-FOUR-CHECK-NEXT:    [[TMP62:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 4
292; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD14:%.*]] = load <4 x float>, ptr [[TMP62]], align 4
293; VF-FOUR-CHECK-NEXT:    [[TMP64:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 8
294; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD15:%.*]] = load <4 x float>, ptr [[TMP64]], align 4
295; VF-FOUR-CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 12
296; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD16:%.*]] = load <4 x float>, ptr [[TMP66]], align 4
297; VF-FOUR-CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 16
298; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD17:%.*]] = load <4 x float>, ptr [[TMP68]], align 4
299; VF-FOUR-CHECK-NEXT:    [[TMP70:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 20
300; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD18:%.*]] = load <4 x float>, ptr [[TMP70]], align 4
301; VF-FOUR-CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 24
302; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD19:%.*]] = load <4 x float>, ptr [[TMP72]], align 4
303; VF-FOUR-CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 28
304; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD20:%.*]] = load <4 x float>, ptr [[TMP74]], align 4
305; VF-FOUR-CHECK-NEXT:    [[TMP76:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 32
306; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD21:%.*]] = load <4 x float>, ptr [[TMP76]], align 4
307; VF-FOUR-CHECK-NEXT:    [[TMP78:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 36
308; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD22:%.*]] = load <4 x float>, ptr [[TMP78]], align 4
309; VF-FOUR-CHECK-NEXT:    [[TMP80:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 40
310; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD23:%.*]] = load <4 x float>, ptr [[TMP80]], align 4
311; VF-FOUR-CHECK-NEXT:    [[TMP82:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 44
312; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD24:%.*]] = load <4 x float>, ptr [[TMP82]], align 4
313; VF-FOUR-CHECK-NEXT:    [[TMP84:%.*]] = fadd fast <4 x float> [[WIDE_LOAD]], [[WIDE_LOAD13]]
314; VF-FOUR-CHECK-NEXT:    [[TMP85:%.*]] = fadd fast <4 x float> [[WIDE_LOAD2]], [[WIDE_LOAD14]]
315; VF-FOUR-CHECK-NEXT:    [[TMP86:%.*]] = fadd fast <4 x float> [[WIDE_LOAD3]], [[WIDE_LOAD15]]
316; VF-FOUR-CHECK-NEXT:    [[TMP87:%.*]] = fadd fast <4 x float> [[WIDE_LOAD4]], [[WIDE_LOAD16]]
317; VF-FOUR-CHECK-NEXT:    [[TMP88:%.*]] = fadd fast <4 x float> [[WIDE_LOAD5]], [[WIDE_LOAD17]]
318; VF-FOUR-CHECK-NEXT:    [[TMP89:%.*]] = fadd fast <4 x float> [[WIDE_LOAD6]], [[WIDE_LOAD18]]
319; VF-FOUR-CHECK-NEXT:    [[TMP90:%.*]] = fadd fast <4 x float> [[WIDE_LOAD7]], [[WIDE_LOAD19]]
320; VF-FOUR-CHECK-NEXT:    [[TMP91:%.*]] = fadd fast <4 x float> [[WIDE_LOAD8]], [[WIDE_LOAD20]]
321; VF-FOUR-CHECK-NEXT:    [[TMP92:%.*]] = fadd fast <4 x float> [[WIDE_LOAD9]], [[WIDE_LOAD21]]
322; VF-FOUR-CHECK-NEXT:    [[TMP93:%.*]] = fadd fast <4 x float> [[WIDE_LOAD10]], [[WIDE_LOAD22]]
323; VF-FOUR-CHECK-NEXT:    [[TMP94:%.*]] = fadd fast <4 x float> [[WIDE_LOAD11]], [[WIDE_LOAD23]]
324; VF-FOUR-CHECK-NEXT:    [[TMP95:%.*]] = fadd fast <4 x float> [[WIDE_LOAD12]], [[WIDE_LOAD24]]
325; VF-FOUR-CHECK-NEXT:    [[TMP96:%.*]] = getelementptr inbounds float, ptr [[AA:%.*]], i64 [[TMP0]]
326; VF-FOUR-CHECK-NEXT:    [[TMP97:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP1]]
327; VF-FOUR-CHECK-NEXT:    [[TMP98:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP2]]
328; VF-FOUR-CHECK-NEXT:    [[TMP99:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP3]]
329; VF-FOUR-CHECK-NEXT:    [[TMP100:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP4]]
330; VF-FOUR-CHECK-NEXT:    [[TMP101:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP5]]
331; VF-FOUR-CHECK-NEXT:    [[TMP102:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP6]]
332; VF-FOUR-CHECK-NEXT:    [[TMP103:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP7]]
333; VF-FOUR-CHECK-NEXT:    [[TMP104:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP8]]
334; VF-FOUR-CHECK-NEXT:    [[TMP105:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP9]]
335; VF-FOUR-CHECK-NEXT:    [[TMP106:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP10]]
336; VF-FOUR-CHECK-NEXT:    [[TMP107:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP11]]
337; VF-FOUR-CHECK-NEXT:    [[TMP108:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 0
338; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP84]], ptr [[TMP108]], align 4
339; VF-FOUR-CHECK-NEXT:    [[TMP110:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 4
340; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP85]], ptr [[TMP110]], align 4
341; VF-FOUR-CHECK-NEXT:    [[TMP112:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 8
342; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP86]], ptr [[TMP112]], align 4
343; VF-FOUR-CHECK-NEXT:    [[TMP114:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 12
344; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP87]], ptr [[TMP114]], align 4
345; VF-FOUR-CHECK-NEXT:    [[TMP116:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 16
346; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP88]], ptr [[TMP116]], align 4
347; VF-FOUR-CHECK-NEXT:    [[TMP118:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 20
348; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP89]], ptr [[TMP118]], align 4
349; VF-FOUR-CHECK-NEXT:    [[TMP120:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 24
350; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP90]], ptr [[TMP120]], align 4
351; VF-FOUR-CHECK-NEXT:    [[TMP122:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 28
352; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP91]], ptr [[TMP122]], align 4
353; VF-FOUR-CHECK-NEXT:    [[TMP124:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 32
354; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP92]], ptr [[TMP124]], align 4
355; VF-FOUR-CHECK-NEXT:    [[TMP126:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 36
356; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP93]], ptr [[TMP126]], align 4
357; VF-FOUR-CHECK-NEXT:    [[TMP128:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 40
358; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP94]], ptr [[TMP128]], align 4
359; VF-FOUR-CHECK-NEXT:    [[TMP130:%.*]] = getelementptr inbounds float, ptr [[TMP96]], i32 44
360; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP95]], ptr [[TMP130]], align 4
361; VF-FOUR-CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 48
362; VF-FOUR-CHECK-NEXT:    [[TMP132:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
363; VF-FOUR-CHECK-NEXT:    br i1 [[TMP132]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
364; VF-FOUR-CHECK:       middle.block:
365; VF-FOUR-CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[WIDE_TRIP_COUNT]], [[N_VEC]]
366; VF-FOUR-CHECK-NEXT:    br i1 [[CMP_N]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
367; VF-FOUR-CHECK:       vec.epilog.iter.check:
368; VF-FOUR-CHECK-NEXT:    [[N_VEC_REMAINING:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_VEC]]
369; VF-FOUR-CHECK-NEXT:    [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_VEC_REMAINING]], 4
370; VF-FOUR-CHECK-NEXT:    br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
371; VF-FOUR-CHECK:       vec.epilog.ph:
372; VF-FOUR-CHECK-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
373; VF-FOUR-CHECK-NEXT:    [[N_MOD_VF25:%.*]] = urem i64 [[WIDE_TRIP_COUNT]], 4
374; VF-FOUR-CHECK-NEXT:    [[N_VEC26:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_MOD_VF25]]
375; VF-FOUR-CHECK-NEXT:    br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
376; VF-FOUR-CHECK:       vec.epilog.vector.body:
377; VF-FOUR-CHECK-NEXT:    [[OFFSET_IDX:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT31:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
378; VF-FOUR-CHECK-NEXT:    [[TMP133:%.*]] = add i64 [[OFFSET_IDX]], 0
379; VF-FOUR-CHECK-NEXT:    [[TMP134:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[TMP133]]
380; VF-FOUR-CHECK-NEXT:    [[TMP135:%.*]] = getelementptr inbounds float, ptr [[TMP134]], i32 0
381; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD29:%.*]] = load <4 x float>, ptr [[TMP135]], align 4
382; VF-FOUR-CHECK-NEXT:    [[TMP137:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[TMP133]]
383; VF-FOUR-CHECK-NEXT:    [[TMP138:%.*]] = getelementptr inbounds float, ptr [[TMP137]], i32 0
384; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD30:%.*]] = load <4 x float>, ptr [[TMP138]], align 4
385; VF-FOUR-CHECK-NEXT:    [[TMP140:%.*]] = fadd fast <4 x float> [[WIDE_LOAD29]], [[WIDE_LOAD30]]
386; VF-FOUR-CHECK-NEXT:    [[TMP141:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[TMP133]]
387; VF-FOUR-CHECK-NEXT:    [[TMP142:%.*]] = getelementptr inbounds float, ptr [[TMP141]], i32 0
388; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP140]], ptr [[TMP142]], align 4
389; VF-FOUR-CHECK-NEXT:    [[INDEX_NEXT31]] = add nuw i64 [[OFFSET_IDX]], 4
390; VF-FOUR-CHECK-NEXT:    [[TMP144:%.*]] = icmp eq i64 [[INDEX_NEXT31]], [[N_VEC26]]
391; VF-FOUR-CHECK-NEXT:    br i1 [[TMP144]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP2:![0-9]+]]
392; VF-FOUR-CHECK:       vec.epilog.middle.block:
393; VF-FOUR-CHECK-NEXT:    [[CMP_N27:%.*]] = icmp eq i64 [[WIDE_TRIP_COUNT]], [[N_VEC26]]
394; VF-FOUR-CHECK-NEXT:    br i1 [[CMP_N27]], label [[FOR_END_LOOPEXIT]], label [[VEC_EPILOG_SCALAR_PH]]
395; VF-FOUR-CHECK:       vec.epilog.scalar.ph:
396; VF-FOUR-CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC26]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[ITER_CHECK]] ]
397; VF-FOUR-CHECK-NEXT:    br label [[FOR_BODY:%.*]]
398; VF-FOUR-CHECK:       for.body:
399; VF-FOUR-CHECK-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
400; VF-FOUR-CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[BB]], i64 [[INDVARS_IV]]
401; VF-FOUR-CHECK-NEXT:    [[TMP145:%.*]] = load float, ptr [[ARRAYIDX]], align 4
402; VF-FOUR-CHECK-NEXT:    [[ARRAYIDX2:%.*]] = getelementptr inbounds float, ptr [[CC]], i64 [[INDVARS_IV]]
403; VF-FOUR-CHECK-NEXT:    [[TMP146:%.*]] = load float, ptr [[ARRAYIDX2]], align 4
404; VF-FOUR-CHECK-NEXT:    [[ADD:%.*]] = fadd fast float [[TMP145]], [[TMP146]]
405; VF-FOUR-CHECK-NEXT:    [[ARRAYIDX4:%.*]] = getelementptr inbounds float, ptr [[AA]], i64 [[INDVARS_IV]]
406; VF-FOUR-CHECK-NEXT:    store float [[ADD]], ptr [[ARRAYIDX4]], align 4
407; VF-FOUR-CHECK-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
408; VF-FOUR-CHECK-NEXT:    [[EXITCOND:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT]], [[WIDE_TRIP_COUNT]]
409; VF-FOUR-CHECK-NEXT:    br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT]], !llvm.loop [[LOOP4:![0-9]+]]
410; VF-FOUR-CHECK:       for.end.loopexit:
411; VF-FOUR-CHECK-NEXT:    br label [[FOR_END]]
412; VF-FOUR-CHECK:       for.end:
413; VF-FOUR-CHECK-NEXT:    ret void
414;
415
416
417entry:
418  %cmp1 = icmp sgt i32 %N, 0
419  br i1 %cmp1, label %for.body.preheader, label %for.end
420
421for.body.preheader:                               ; preds = %entry
422  %wide.trip.count = zext i32 %N to i64
423  br label %for.body
424
425for.body:                                         ; preds = %for.body.preheader, %for.body
426  %indvars.iv = phi i64 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
427  %arrayidx = getelementptr inbounds float, ptr %bb, i64 %indvars.iv
428  %0 = load float, ptr %arrayidx, align 4
429  %arrayidx2 = getelementptr inbounds float, ptr %cc, i64 %indvars.iv
430  %1 = load float, ptr %arrayidx2, align 4
431  %add = fadd fast float %0, %1
432  %arrayidx4 = getelementptr inbounds float, ptr %aa, i64 %indvars.iv
433  store float %add, ptr %arrayidx4, align 4
434  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
435  %exitcond = icmp ne i64 %indvars.iv.next, %wide.trip.count
436  br i1 %exitcond, label %for.body, label %for.end.loopexit
437
438for.end.loopexit:                                 ; preds = %for.body
439  br label %for.end
440
441for.end:                                          ; preds = %for.end.loopexit, %entry
442  ret void
443}
444
445define dso_local signext i32 @f2(ptr noalias %A, ptr noalias %B, i32 signext %n) #0 {
446; VF-TWO-CHECK-LABEL: @f2(
447; VF-TWO-CHECK-NEXT:  entry:
448; VF-TWO-CHECK-NEXT:    [[CMP1:%.*]] = icmp sgt i32 [[N:%.*]], 1
449; VF-TWO-CHECK-NEXT:    br i1 [[CMP1]], label [[ITER_CHECK:%.*]], label [[FOR_END:%.*]]
450; VF-TWO-CHECK:       iter.check:
451; VF-TWO-CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[N]], -1
452; VF-TWO-CHECK-NEXT:    [[WIDE_TRIP_COUNT:%.*]] = zext i32 [[TMP0]] to i64
453; VF-TWO-CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[WIDE_TRIP_COUNT]], 2
454; VF-TWO-CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_SCEVCHECK:%.*]]
455; VF-TWO-CHECK:       vector.scevcheck:
456; VF-TWO-CHECK-NEXT:    [[TMP1:%.*]] = add nsw i64 [[WIDE_TRIP_COUNT]], -1
457; VF-TWO-CHECK-NEXT:    [[TMP2:%.*]] = trunc i64 [[TMP1]] to i32
458; VF-TWO-CHECK-NEXT:    [[MUL:%.*]] = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 1, i32 [[TMP2]])
459; VF-TWO-CHECK-NEXT:    [[MUL_RESULT:%.*]] = extractvalue { i32, i1 } [[MUL]], 0
460; VF-TWO-CHECK-NEXT:    [[MUL_OVERFLOW:%.*]] = extractvalue { i32, i1 } [[MUL]], 1
461; VF-TWO-CHECK-NEXT:    [[TMP3:%.*]] = sub i32 [[TMP0]], [[MUL_RESULT]]
462; VF-TWO-CHECK-NEXT:    [[TMP4:%.*]] = icmp sgt i32 [[TMP3]], [[TMP0]]
463; VF-TWO-CHECK-NEXT:    [[TMP5:%.*]] = or i1 [[TMP4]], [[MUL_OVERFLOW]]
464; VF-TWO-CHECK-NEXT:    [[TMP6:%.*]] = icmp ugt i64 [[TMP1]], 4294967295
465; VF-TWO-CHECK-NEXT:    [[TMP7:%.*]] = or i1 [[TMP5]], [[TMP6]]
466; VF-TWO-CHECK-NEXT:    br i1 [[TMP7]], label [[VEC_EPILOG_SCALAR_PH]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
467; VF-TWO-CHECK:       vector.main.loop.iter.check:
468; VF-TWO-CHECK-NEXT:    [[MIN_ITERS_CHECK1:%.*]] = icmp ult i64 [[WIDE_TRIP_COUNT]], 32
469; VF-TWO-CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
470; VF-TWO-CHECK:       vector.ph:
471; VF-TWO-CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[WIDE_TRIP_COUNT]], 32
472; VF-TWO-CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_MOD_VF]]
473; VF-TWO-CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
474; VF-TWO-CHECK:       vector.body:
475; VF-TWO-CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
476; VF-TWO-CHECK-NEXT:    [[OFFSET_IDX:%.*]] = trunc i64 [[INDEX]] to i32
477; VF-TWO-CHECK-NEXT:    [[TMP8:%.*]] = add i32 [[OFFSET_IDX]], 0
478; VF-TWO-CHECK-NEXT:    [[TMP9:%.*]] = add i32 [[OFFSET_IDX]], 4
479; VF-TWO-CHECK-NEXT:    [[TMP10:%.*]] = add i32 [[OFFSET_IDX]], 8
480; VF-TWO-CHECK-NEXT:    [[TMP11:%.*]] = add i32 [[OFFSET_IDX]], 12
481; VF-TWO-CHECK-NEXT:    [[TMP12:%.*]] = add i32 [[OFFSET_IDX]], 16
482; VF-TWO-CHECK-NEXT:    [[TMP13:%.*]] = add i32 [[OFFSET_IDX]], 20
483; VF-TWO-CHECK-NEXT:    [[TMP14:%.*]] = add i32 [[OFFSET_IDX]], 24
484; VF-TWO-CHECK-NEXT:    [[TMP15:%.*]] = add i32 [[OFFSET_IDX]], 28
485; VF-TWO-CHECK-NEXT:    [[TMP16:%.*]] = add i64 [[INDEX]], 0
486; VF-TWO-CHECK-NEXT:    [[TMP17:%.*]] = add i64 [[INDEX]], 4
487; VF-TWO-CHECK-NEXT:    [[TMP18:%.*]] = add i64 [[INDEX]], 8
488; VF-TWO-CHECK-NEXT:    [[TMP19:%.*]] = add i64 [[INDEX]], 12
489; VF-TWO-CHECK-NEXT:    [[TMP20:%.*]] = add i64 [[INDEX]], 16
490; VF-TWO-CHECK-NEXT:    [[TMP21:%.*]] = add i64 [[INDEX]], 20
491; VF-TWO-CHECK-NEXT:    [[TMP22:%.*]] = add i64 [[INDEX]], 24
492; VF-TWO-CHECK-NEXT:    [[TMP23:%.*]] = add i64 [[INDEX]], 28
493; VF-TWO-CHECK-NEXT:    [[TMP24:%.*]] = xor i32 [[TMP8]], -1
494; VF-TWO-CHECK-NEXT:    [[TMP25:%.*]] = xor i32 [[TMP9]], -1
495; VF-TWO-CHECK-NEXT:    [[TMP26:%.*]] = xor i32 [[TMP10]], -1
496; VF-TWO-CHECK-NEXT:    [[TMP27:%.*]] = xor i32 [[TMP11]], -1
497; VF-TWO-CHECK-NEXT:    [[TMP28:%.*]] = xor i32 [[TMP12]], -1
498; VF-TWO-CHECK-NEXT:    [[TMP29:%.*]] = xor i32 [[TMP13]], -1
499; VF-TWO-CHECK-NEXT:    [[TMP30:%.*]] = xor i32 [[TMP14]], -1
500; VF-TWO-CHECK-NEXT:    [[TMP31:%.*]] = xor i32 [[TMP15]], -1
501; VF-TWO-CHECK-NEXT:    [[TMP32:%.*]] = add i32 [[TMP24]], [[N]]
502; VF-TWO-CHECK-NEXT:    [[TMP33:%.*]] = add i32 [[TMP25]], [[N]]
503; VF-TWO-CHECK-NEXT:    [[TMP34:%.*]] = add i32 [[TMP26]], [[N]]
504; VF-TWO-CHECK-NEXT:    [[TMP35:%.*]] = add i32 [[TMP27]], [[N]]
505; VF-TWO-CHECK-NEXT:    [[TMP36:%.*]] = add i32 [[TMP28]], [[N]]
506; VF-TWO-CHECK-NEXT:    [[TMP37:%.*]] = add i32 [[TMP29]], [[N]]
507; VF-TWO-CHECK-NEXT:    [[TMP38:%.*]] = add i32 [[TMP30]], [[N]]
508; VF-TWO-CHECK-NEXT:    [[TMP39:%.*]] = add i32 [[TMP31]], [[N]]
509; VF-TWO-CHECK-NEXT:    [[TMP40:%.*]] = sext i32 [[TMP32]] to i64
510; VF-TWO-CHECK-NEXT:    [[TMP41:%.*]] = sext i32 [[TMP33]] to i64
511; VF-TWO-CHECK-NEXT:    [[TMP42:%.*]] = sext i32 [[TMP34]] to i64
512; VF-TWO-CHECK-NEXT:    [[TMP43:%.*]] = sext i32 [[TMP35]] to i64
513; VF-TWO-CHECK-NEXT:    [[TMP44:%.*]] = sext i32 [[TMP36]] to i64
514; VF-TWO-CHECK-NEXT:    [[TMP45:%.*]] = sext i32 [[TMP37]] to i64
515; VF-TWO-CHECK-NEXT:    [[TMP46:%.*]] = sext i32 [[TMP38]] to i64
516; VF-TWO-CHECK-NEXT:    [[TMP47:%.*]] = sext i32 [[TMP39]] to i64
517; VF-TWO-CHECK-NEXT:    [[TMP48:%.*]] = getelementptr inbounds float, ptr [[B:%.*]], i64 [[TMP40]]
518; VF-TWO-CHECK-NEXT:    [[TMP49:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP41]]
519; VF-TWO-CHECK-NEXT:    [[TMP50:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP42]]
520; VF-TWO-CHECK-NEXT:    [[TMP51:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP43]]
521; VF-TWO-CHECK-NEXT:    [[TMP52:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP44]]
522; VF-TWO-CHECK-NEXT:    [[TMP53:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP45]]
523; VF-TWO-CHECK-NEXT:    [[TMP54:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP46]]
524; VF-TWO-CHECK-NEXT:    [[TMP55:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP47]]
525; VF-TWO-CHECK-NEXT:    [[TMP56:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 0
526; VF-TWO-CHECK-NEXT:    [[TMP57:%.*]] = getelementptr inbounds float, ptr [[TMP56]], i32 -3
527; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP57]], align 4
528; VF-TWO-CHECK-NEXT:    [[REVERSE:%.*]] = shufflevector <4 x float> [[WIDE_LOAD]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
529; VF-TWO-CHECK-NEXT:    [[TMP59:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -4
530; VF-TWO-CHECK-NEXT:    [[TMP60:%.*]] = getelementptr inbounds float, ptr [[TMP59]], i32 -3
531; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD2:%.*]] = load <4 x float>, ptr [[TMP60]], align 4
532; VF-TWO-CHECK-NEXT:    [[REVERSE3:%.*]] = shufflevector <4 x float> [[WIDE_LOAD2]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
533; VF-TWO-CHECK-NEXT:    [[TMP62:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -8
534; VF-TWO-CHECK-NEXT:    [[TMP63:%.*]] = getelementptr inbounds float, ptr [[TMP62]], i32 -3
535; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x float>, ptr [[TMP63]], align 4
536; VF-TWO-CHECK-NEXT:    [[REVERSE5:%.*]] = shufflevector <4 x float> [[WIDE_LOAD4]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
537; VF-TWO-CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -12
538; VF-TWO-CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds float, ptr [[TMP65]], i32 -3
539; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x float>, ptr [[TMP66]], align 4
540; VF-TWO-CHECK-NEXT:    [[REVERSE7:%.*]] = shufflevector <4 x float> [[WIDE_LOAD6]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
541; VF-TWO-CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -16
542; VF-TWO-CHECK-NEXT:    [[TMP69:%.*]] = getelementptr inbounds float, ptr [[TMP68]], i32 -3
543; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD8:%.*]] = load <4 x float>, ptr [[TMP69]], align 4
544; VF-TWO-CHECK-NEXT:    [[REVERSE9:%.*]] = shufflevector <4 x float> [[WIDE_LOAD8]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
545; VF-TWO-CHECK-NEXT:    [[TMP71:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -20
546; VF-TWO-CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds float, ptr [[TMP71]], i32 -3
547; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD10:%.*]] = load <4 x float>, ptr [[TMP72]], align 4
548; VF-TWO-CHECK-NEXT:    [[REVERSE11:%.*]] = shufflevector <4 x float> [[WIDE_LOAD10]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
549; VF-TWO-CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -24
550; VF-TWO-CHECK-NEXT:    [[TMP75:%.*]] = getelementptr inbounds float, ptr [[TMP74]], i32 -3
551; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD12:%.*]] = load <4 x float>, ptr [[TMP75]], align 4
552; VF-TWO-CHECK-NEXT:    [[REVERSE13:%.*]] = shufflevector <4 x float> [[WIDE_LOAD12]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
553; VF-TWO-CHECK-NEXT:    [[TMP77:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -28
554; VF-TWO-CHECK-NEXT:    [[TMP78:%.*]] = getelementptr inbounds float, ptr [[TMP77]], i32 -3
555; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD14:%.*]] = load <4 x float>, ptr [[TMP78]], align 4
556; VF-TWO-CHECK-NEXT:    [[REVERSE15:%.*]] = shufflevector <4 x float> [[WIDE_LOAD14]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
557; VF-TWO-CHECK-NEXT:    [[TMP80:%.*]] = fadd fast <4 x float> [[REVERSE]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
558; VF-TWO-CHECK-NEXT:    [[TMP81:%.*]] = fadd fast <4 x float> [[REVERSE3]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
559; VF-TWO-CHECK-NEXT:    [[TMP82:%.*]] = fadd fast <4 x float> [[REVERSE5]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
560; VF-TWO-CHECK-NEXT:    [[TMP83:%.*]] = fadd fast <4 x float> [[REVERSE7]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
561; VF-TWO-CHECK-NEXT:    [[TMP84:%.*]] = fadd fast <4 x float> [[REVERSE9]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
562; VF-TWO-CHECK-NEXT:    [[TMP85:%.*]] = fadd fast <4 x float> [[REVERSE11]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
563; VF-TWO-CHECK-NEXT:    [[TMP86:%.*]] = fadd fast <4 x float> [[REVERSE13]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
564; VF-TWO-CHECK-NEXT:    [[TMP87:%.*]] = fadd fast <4 x float> [[REVERSE15]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
565; VF-TWO-CHECK-NEXT:    [[TMP88:%.*]] = getelementptr inbounds float, ptr [[A:%.*]], i64 [[TMP16]]
566; VF-TWO-CHECK-NEXT:    [[TMP89:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP17]]
567; VF-TWO-CHECK-NEXT:    [[TMP90:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP18]]
568; VF-TWO-CHECK-NEXT:    [[TMP91:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP19]]
569; VF-TWO-CHECK-NEXT:    [[TMP92:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP20]]
570; VF-TWO-CHECK-NEXT:    [[TMP93:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP21]]
571; VF-TWO-CHECK-NEXT:    [[TMP94:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP22]]
572; VF-TWO-CHECK-NEXT:    [[TMP95:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP23]]
573; VF-TWO-CHECK-NEXT:    [[TMP96:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 0
574; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP80]], ptr [[TMP96]], align 4
575; VF-TWO-CHECK-NEXT:    [[TMP98:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 4
576; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP81]], ptr [[TMP98]], align 4
577; VF-TWO-CHECK-NEXT:    [[TMP100:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 8
578; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP82]], ptr [[TMP100]], align 4
579; VF-TWO-CHECK-NEXT:    [[TMP102:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 12
580; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP83]], ptr [[TMP102]], align 4
581; VF-TWO-CHECK-NEXT:    [[TMP104:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 16
582; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP84]], ptr [[TMP104]], align 4
583; VF-TWO-CHECK-NEXT:    [[TMP106:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 20
584; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP85]], ptr [[TMP106]], align 4
585; VF-TWO-CHECK-NEXT:    [[TMP108:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 24
586; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP86]], ptr [[TMP108]], align 4
587; VF-TWO-CHECK-NEXT:    [[TMP110:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 28
588; VF-TWO-CHECK-NEXT:    store <4 x float> [[TMP87]], ptr [[TMP110]], align 4
589; VF-TWO-CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 32
590; VF-TWO-CHECK-NEXT:    [[TMP112:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
591; VF-TWO-CHECK-NEXT:    br i1 [[TMP112]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
592; VF-TWO-CHECK:       middle.block:
593; VF-TWO-CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[WIDE_TRIP_COUNT]], [[N_VEC]]
594; VF-TWO-CHECK-NEXT:    br i1 [[CMP_N]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
595; VF-TWO-CHECK:       vec.epilog.iter.check:
596; VF-TWO-CHECK-NEXT:    [[IND_END19:%.*]] = trunc i64 [[N_VEC]] to i32
597; VF-TWO-CHECK-NEXT:    [[N_VEC_REMAINING:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_VEC]]
598; VF-TWO-CHECK-NEXT:    [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_VEC_REMAINING]], 2
599; VF-TWO-CHECK-NEXT:    br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
600; VF-TWO-CHECK:       vec.epilog.ph:
601; VF-TWO-CHECK-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
602; VF-TWO-CHECK-NEXT:    [[N_MOD_VF16:%.*]] = urem i64 [[WIDE_TRIP_COUNT]], 2
603; VF-TWO-CHECK-NEXT:    [[N_VEC17:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_MOD_VF16]]
604; VF-TWO-CHECK-NEXT:    [[IND_END:%.*]] = trunc i64 [[N_VEC17]] to i32
605; VF-TWO-CHECK-NEXT:    br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
606; VF-TWO-CHECK:       vec.epilog.vector.body:
607; VF-TWO-CHECK-NEXT:    [[OFFSET_IDX23:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT26:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
608; VF-TWO-CHECK-NEXT:    [[OFFSET_IDX22:%.*]] = trunc i64 [[OFFSET_IDX23]] to i32
609; VF-TWO-CHECK-NEXT:    [[TMP113:%.*]] = add i32 [[OFFSET_IDX22]], 0
610; VF-TWO-CHECK-NEXT:    [[TMP114:%.*]] = add i64 [[OFFSET_IDX23]], 0
611; VF-TWO-CHECK-NEXT:    [[TMP115:%.*]] = xor i32 [[TMP113]], -1
612; VF-TWO-CHECK-NEXT:    [[TMP116:%.*]] = add i32 [[TMP115]], [[N]]
613; VF-TWO-CHECK-NEXT:    [[TMP117:%.*]] = sext i32 [[TMP116]] to i64
614; VF-TWO-CHECK-NEXT:    [[TMP118:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP117]]
615; VF-TWO-CHECK-NEXT:    [[TMP119:%.*]] = getelementptr inbounds float, ptr [[TMP118]], i32 0
616; VF-TWO-CHECK-NEXT:    [[TMP120:%.*]] = getelementptr inbounds float, ptr [[TMP119]], i32 -1
617; VF-TWO-CHECK-NEXT:    [[WIDE_LOAD24:%.*]] = load <2 x float>, ptr [[TMP120]], align 4
618; VF-TWO-CHECK-NEXT:    [[REVERSE25:%.*]] = shufflevector <2 x float> [[WIDE_LOAD24]], <2 x float> poison, <2 x i32> <i32 1, i32 0>
619; VF-TWO-CHECK-NEXT:    [[TMP122:%.*]] = fadd fast <2 x float> [[REVERSE25]], <float 1.000000e+00, float 1.000000e+00>
620; VF-TWO-CHECK-NEXT:    [[TMP123:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP114]]
621; VF-TWO-CHECK-NEXT:    [[TMP124:%.*]] = getelementptr inbounds float, ptr [[TMP123]], i32 0
622; VF-TWO-CHECK-NEXT:    store <2 x float> [[TMP122]], ptr [[TMP124]], align 4
623; VF-TWO-CHECK-NEXT:    [[INDEX_NEXT26]] = add nuw i64 [[OFFSET_IDX23]], 2
624; VF-TWO-CHECK-NEXT:    [[TMP126:%.*]] = icmp eq i64 [[INDEX_NEXT26]], [[N_VEC17]]
625; VF-TWO-CHECK-NEXT:    br i1 [[TMP126]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
626; VF-TWO-CHECK:       vec.epilog.middle.block:
627; VF-TWO-CHECK-NEXT:    [[CMP_N20:%.*]] = icmp eq i64 [[WIDE_TRIP_COUNT]], [[N_VEC17]]
628; VF-TWO-CHECK-NEXT:    br i1 [[CMP_N20]], label [[FOR_END_LOOPEXIT]], label [[VEC_EPILOG_SCALAR_PH]]
629; VF-TWO-CHECK:       vec.epilog.scalar.ph:
630; VF-TWO-CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC17]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_SCEVCHECK]] ], [ 0, [[ITER_CHECK]] ]
631; VF-TWO-CHECK-NEXT:    [[BC_RESUME_VAL18:%.*]] = phi i32 [ [[IND_END]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END19]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_SCEVCHECK]] ], [ 0, [[ITER_CHECK]] ]
632; VF-TWO-CHECK-NEXT:    br label [[FOR_BODY:%.*]]
633; VF-TWO-CHECK:       for.body:
634; VF-TWO-CHECK-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
635; VF-TWO-CHECK-NEXT:    [[I_014:%.*]] = phi i32 [ [[BC_RESUME_VAL18]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INC:%.*]], [[FOR_BODY]] ]
636; VF-TWO-CHECK-NEXT:    [[TMP127:%.*]] = xor i32 [[I_014]], -1
637; VF-TWO-CHECK-NEXT:    [[SUB2:%.*]] = add i32 [[TMP127]], [[N]]
638; VF-TWO-CHECK-NEXT:    [[IDXPROM:%.*]] = sext i32 [[SUB2]] to i64
639; VF-TWO-CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[IDXPROM]]
640; VF-TWO-CHECK-NEXT:    [[TMP128:%.*]] = load float, ptr [[ARRAYIDX]], align 4
641; VF-TWO-CHECK-NEXT:    [[CONV3:%.*]] = fadd fast float [[TMP128]], 1.000000e+00
642; VF-TWO-CHECK-NEXT:    [[ARRAYIDX5:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[INDVARS_IV]]
643; VF-TWO-CHECK-NEXT:    store float [[CONV3]], ptr [[ARRAYIDX5]], align 4
644; VF-TWO-CHECK-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
645; VF-TWO-CHECK-NEXT:    [[INC]] = add nuw nsw i32 [[I_014]], 1
646; VF-TWO-CHECK-NEXT:    [[EXITCOND:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT]], [[WIDE_TRIP_COUNT]]
647; VF-TWO-CHECK-NEXT:    br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT]], !llvm.loop [[LOOPID_MS_CM:![0-9]+]]
648; VF-TWO-CHECK:       for.end.loopexit:
649; VF-TWO-CHECK-NEXT:    br label [[FOR_END]]
650; VF-TWO-CHECK:       for.end:
651; VF-TWO-CHECK-NEXT:    ret i32 0
652;
653; VF-FOUR-CHECK-LABEL: @f2(
654; VF-FOUR-CHECK-NEXT:  entry:
655; VF-FOUR-CHECK-NEXT:    [[CMP1:%.*]] = icmp sgt i32 [[N:%.*]], 1
656; VF-FOUR-CHECK-NEXT:    br i1 [[CMP1]], label [[ITER_CHECK:%.*]], label [[FOR_END:%.*]]
657; VF-FOUR-CHECK:       iter.check:
658; VF-FOUR-CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[N]], -1
659; VF-FOUR-CHECK-NEXT:    [[WIDE_TRIP_COUNT:%.*]] = zext i32 [[TMP0]] to i64
660; VF-FOUR-CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[WIDE_TRIP_COUNT]], 4
661; VF-FOUR-CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_SCEVCHECK:%.*]]
662; VF-FOUR-CHECK:       vector.scevcheck:
663; VF-FOUR-CHECK-NEXT:    [[TMP1:%.*]] = add nsw i64 [[WIDE_TRIP_COUNT]], -1
664; VF-FOUR-CHECK-NEXT:    [[TMP2:%.*]] = trunc i64 [[TMP1]] to i32
665; VF-FOUR-CHECK-NEXT:    [[MUL:%.*]] = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 1, i32 [[TMP2]])
666; VF-FOUR-CHECK-NEXT:    [[MUL_RESULT:%.*]] = extractvalue { i32, i1 } [[MUL]], 0
667; VF-FOUR-CHECK-NEXT:    [[MUL_OVERFLOW:%.*]] = extractvalue { i32, i1 } [[MUL]], 1
668; VF-FOUR-CHECK-NEXT:    [[TMP3:%.*]] = sub i32 [[TMP0]], [[MUL_RESULT]]
669; VF-FOUR-CHECK-NEXT:    [[TMP4:%.*]] = icmp sgt i32 [[TMP3]], [[TMP0]]
670; VF-FOUR-CHECK-NEXT:    [[TMP5:%.*]] = or i1 [[TMP4]], [[MUL_OVERFLOW]]
671; VF-FOUR-CHECK-NEXT:    [[TMP6:%.*]] = icmp ugt i64 [[TMP1]], 4294967295
672; VF-FOUR-CHECK-NEXT:    [[TMP7:%.*]] = or i1 [[TMP5]], [[TMP6]]
673; VF-FOUR-CHECK-NEXT:    br i1 [[TMP7]], label [[VEC_EPILOG_SCALAR_PH]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
674; VF-FOUR-CHECK:       vector.main.loop.iter.check:
675; VF-FOUR-CHECK-NEXT:    [[MIN_ITERS_CHECK1:%.*]] = icmp ult i64 [[WIDE_TRIP_COUNT]], 32
676; VF-FOUR-CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
677; VF-FOUR-CHECK:       vector.ph:
678; VF-FOUR-CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[WIDE_TRIP_COUNT]], 32
679; VF-FOUR-CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_MOD_VF]]
680; VF-FOUR-CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
681; VF-FOUR-CHECK:       vector.body:
682; VF-FOUR-CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
683; VF-FOUR-CHECK-NEXT:    [[OFFSET_IDX:%.*]] = trunc i64 [[INDEX]] to i32
684; VF-FOUR-CHECK-NEXT:    [[TMP8:%.*]] = add i32 [[OFFSET_IDX]], 0
685; VF-FOUR-CHECK-NEXT:    [[TMP9:%.*]] = add i32 [[OFFSET_IDX]], 4
686; VF-FOUR-CHECK-NEXT:    [[TMP10:%.*]] = add i32 [[OFFSET_IDX]], 8
687; VF-FOUR-CHECK-NEXT:    [[TMP11:%.*]] = add i32 [[OFFSET_IDX]], 12
688; VF-FOUR-CHECK-NEXT:    [[TMP12:%.*]] = add i32 [[OFFSET_IDX]], 16
689; VF-FOUR-CHECK-NEXT:    [[TMP13:%.*]] = add i32 [[OFFSET_IDX]], 20
690; VF-FOUR-CHECK-NEXT:    [[TMP14:%.*]] = add i32 [[OFFSET_IDX]], 24
691; VF-FOUR-CHECK-NEXT:    [[TMP15:%.*]] = add i32 [[OFFSET_IDX]], 28
692; VF-FOUR-CHECK-NEXT:    [[TMP16:%.*]] = add i64 [[INDEX]], 0
693; VF-FOUR-CHECK-NEXT:    [[TMP17:%.*]] = add i64 [[INDEX]], 4
694; VF-FOUR-CHECK-NEXT:    [[TMP18:%.*]] = add i64 [[INDEX]], 8
695; VF-FOUR-CHECK-NEXT:    [[TMP19:%.*]] = add i64 [[INDEX]], 12
696; VF-FOUR-CHECK-NEXT:    [[TMP20:%.*]] = add i64 [[INDEX]], 16
697; VF-FOUR-CHECK-NEXT:    [[TMP21:%.*]] = add i64 [[INDEX]], 20
698; VF-FOUR-CHECK-NEXT:    [[TMP22:%.*]] = add i64 [[INDEX]], 24
699; VF-FOUR-CHECK-NEXT:    [[TMP23:%.*]] = add i64 [[INDEX]], 28
700; VF-FOUR-CHECK-NEXT:    [[TMP24:%.*]] = xor i32 [[TMP8]], -1
701; VF-FOUR-CHECK-NEXT:    [[TMP25:%.*]] = xor i32 [[TMP9]], -1
702; VF-FOUR-CHECK-NEXT:    [[TMP26:%.*]] = xor i32 [[TMP10]], -1
703; VF-FOUR-CHECK-NEXT:    [[TMP27:%.*]] = xor i32 [[TMP11]], -1
704; VF-FOUR-CHECK-NEXT:    [[TMP28:%.*]] = xor i32 [[TMP12]], -1
705; VF-FOUR-CHECK-NEXT:    [[TMP29:%.*]] = xor i32 [[TMP13]], -1
706; VF-FOUR-CHECK-NEXT:    [[TMP30:%.*]] = xor i32 [[TMP14]], -1
707; VF-FOUR-CHECK-NEXT:    [[TMP31:%.*]] = xor i32 [[TMP15]], -1
708; VF-FOUR-CHECK-NEXT:    [[TMP32:%.*]] = add i32 [[TMP24]], [[N]]
709; VF-FOUR-CHECK-NEXT:    [[TMP33:%.*]] = add i32 [[TMP25]], [[N]]
710; VF-FOUR-CHECK-NEXT:    [[TMP34:%.*]] = add i32 [[TMP26]], [[N]]
711; VF-FOUR-CHECK-NEXT:    [[TMP35:%.*]] = add i32 [[TMP27]], [[N]]
712; VF-FOUR-CHECK-NEXT:    [[TMP36:%.*]] = add i32 [[TMP28]], [[N]]
713; VF-FOUR-CHECK-NEXT:    [[TMP37:%.*]] = add i32 [[TMP29]], [[N]]
714; VF-FOUR-CHECK-NEXT:    [[TMP38:%.*]] = add i32 [[TMP30]], [[N]]
715; VF-FOUR-CHECK-NEXT:    [[TMP39:%.*]] = add i32 [[TMP31]], [[N]]
716; VF-FOUR-CHECK-NEXT:    [[TMP40:%.*]] = sext i32 [[TMP32]] to i64
717; VF-FOUR-CHECK-NEXT:    [[TMP41:%.*]] = sext i32 [[TMP33]] to i64
718; VF-FOUR-CHECK-NEXT:    [[TMP42:%.*]] = sext i32 [[TMP34]] to i64
719; VF-FOUR-CHECK-NEXT:    [[TMP43:%.*]] = sext i32 [[TMP35]] to i64
720; VF-FOUR-CHECK-NEXT:    [[TMP44:%.*]] = sext i32 [[TMP36]] to i64
721; VF-FOUR-CHECK-NEXT:    [[TMP45:%.*]] = sext i32 [[TMP37]] to i64
722; VF-FOUR-CHECK-NEXT:    [[TMP46:%.*]] = sext i32 [[TMP38]] to i64
723; VF-FOUR-CHECK-NEXT:    [[TMP47:%.*]] = sext i32 [[TMP39]] to i64
724; VF-FOUR-CHECK-NEXT:    [[TMP48:%.*]] = getelementptr inbounds float, ptr [[B:%.*]], i64 [[TMP40]]
725; VF-FOUR-CHECK-NEXT:    [[TMP49:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP41]]
726; VF-FOUR-CHECK-NEXT:    [[TMP50:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP42]]
727; VF-FOUR-CHECK-NEXT:    [[TMP51:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP43]]
728; VF-FOUR-CHECK-NEXT:    [[TMP52:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP44]]
729; VF-FOUR-CHECK-NEXT:    [[TMP53:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP45]]
730; VF-FOUR-CHECK-NEXT:    [[TMP54:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP46]]
731; VF-FOUR-CHECK-NEXT:    [[TMP55:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP47]]
732; VF-FOUR-CHECK-NEXT:    [[TMP56:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 0
733; VF-FOUR-CHECK-NEXT:    [[TMP57:%.*]] = getelementptr inbounds float, ptr [[TMP56]], i32 -3
734; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP57]], align 4
735; VF-FOUR-CHECK-NEXT:    [[REVERSE:%.*]] = shufflevector <4 x float> [[WIDE_LOAD]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
736; VF-FOUR-CHECK-NEXT:    [[TMP59:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -4
737; VF-FOUR-CHECK-NEXT:    [[TMP60:%.*]] = getelementptr inbounds float, ptr [[TMP59]], i32 -3
738; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD2:%.*]] = load <4 x float>, ptr [[TMP60]], align 4
739; VF-FOUR-CHECK-NEXT:    [[REVERSE3:%.*]] = shufflevector <4 x float> [[WIDE_LOAD2]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
740; VF-FOUR-CHECK-NEXT:    [[TMP62:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -8
741; VF-FOUR-CHECK-NEXT:    [[TMP63:%.*]] = getelementptr inbounds float, ptr [[TMP62]], i32 -3
742; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD4:%.*]] = load <4 x float>, ptr [[TMP63]], align 4
743; VF-FOUR-CHECK-NEXT:    [[REVERSE5:%.*]] = shufflevector <4 x float> [[WIDE_LOAD4]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
744; VF-FOUR-CHECK-NEXT:    [[TMP65:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -12
745; VF-FOUR-CHECK-NEXT:    [[TMP66:%.*]] = getelementptr inbounds float, ptr [[TMP65]], i32 -3
746; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD6:%.*]] = load <4 x float>, ptr [[TMP66]], align 4
747; VF-FOUR-CHECK-NEXT:    [[REVERSE7:%.*]] = shufflevector <4 x float> [[WIDE_LOAD6]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
748; VF-FOUR-CHECK-NEXT:    [[TMP68:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -16
749; VF-FOUR-CHECK-NEXT:    [[TMP69:%.*]] = getelementptr inbounds float, ptr [[TMP68]], i32 -3
750; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD8:%.*]] = load <4 x float>, ptr [[TMP69]], align 4
751; VF-FOUR-CHECK-NEXT:    [[REVERSE9:%.*]] = shufflevector <4 x float> [[WIDE_LOAD8]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
752; VF-FOUR-CHECK-NEXT:    [[TMP71:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -20
753; VF-FOUR-CHECK-NEXT:    [[TMP72:%.*]] = getelementptr inbounds float, ptr [[TMP71]], i32 -3
754; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD10:%.*]] = load <4 x float>, ptr [[TMP72]], align 4
755; VF-FOUR-CHECK-NEXT:    [[REVERSE11:%.*]] = shufflevector <4 x float> [[WIDE_LOAD10]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
756; VF-FOUR-CHECK-NEXT:    [[TMP74:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -24
757; VF-FOUR-CHECK-NEXT:    [[TMP75:%.*]] = getelementptr inbounds float, ptr [[TMP74]], i32 -3
758; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD12:%.*]] = load <4 x float>, ptr [[TMP75]], align 4
759; VF-FOUR-CHECK-NEXT:    [[REVERSE13:%.*]] = shufflevector <4 x float> [[WIDE_LOAD12]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
760; VF-FOUR-CHECK-NEXT:    [[TMP77:%.*]] = getelementptr inbounds float, ptr [[TMP48]], i32 -28
761; VF-FOUR-CHECK-NEXT:    [[TMP78:%.*]] = getelementptr inbounds float, ptr [[TMP77]], i32 -3
762; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD14:%.*]] = load <4 x float>, ptr [[TMP78]], align 4
763; VF-FOUR-CHECK-NEXT:    [[REVERSE15:%.*]] = shufflevector <4 x float> [[WIDE_LOAD14]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
764; VF-FOUR-CHECK-NEXT:    [[TMP80:%.*]] = fadd fast <4 x float> [[REVERSE]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
765; VF-FOUR-CHECK-NEXT:    [[TMP81:%.*]] = fadd fast <4 x float> [[REVERSE3]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
766; VF-FOUR-CHECK-NEXT:    [[TMP82:%.*]] = fadd fast <4 x float> [[REVERSE5]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
767; VF-FOUR-CHECK-NEXT:    [[TMP83:%.*]] = fadd fast <4 x float> [[REVERSE7]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
768; VF-FOUR-CHECK-NEXT:    [[TMP84:%.*]] = fadd fast <4 x float> [[REVERSE9]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
769; VF-FOUR-CHECK-NEXT:    [[TMP85:%.*]] = fadd fast <4 x float> [[REVERSE11]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
770; VF-FOUR-CHECK-NEXT:    [[TMP86:%.*]] = fadd fast <4 x float> [[REVERSE13]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
771; VF-FOUR-CHECK-NEXT:    [[TMP87:%.*]] = fadd fast <4 x float> [[REVERSE15]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
772; VF-FOUR-CHECK-NEXT:    [[TMP88:%.*]] = getelementptr inbounds float, ptr [[A:%.*]], i64 [[TMP16]]
773; VF-FOUR-CHECK-NEXT:    [[TMP89:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP17]]
774; VF-FOUR-CHECK-NEXT:    [[TMP90:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP18]]
775; VF-FOUR-CHECK-NEXT:    [[TMP91:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP19]]
776; VF-FOUR-CHECK-NEXT:    [[TMP92:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP20]]
777; VF-FOUR-CHECK-NEXT:    [[TMP93:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP21]]
778; VF-FOUR-CHECK-NEXT:    [[TMP94:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP22]]
779; VF-FOUR-CHECK-NEXT:    [[TMP95:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP23]]
780; VF-FOUR-CHECK-NEXT:    [[TMP96:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 0
781; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP80]], ptr [[TMP96]], align 4
782; VF-FOUR-CHECK-NEXT:    [[TMP98:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 4
783; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP81]], ptr [[TMP98]], align 4
784; VF-FOUR-CHECK-NEXT:    [[TMP100:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 8
785; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP82]], ptr [[TMP100]], align 4
786; VF-FOUR-CHECK-NEXT:    [[TMP102:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 12
787; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP83]], ptr [[TMP102]], align 4
788; VF-FOUR-CHECK-NEXT:    [[TMP104:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 16
789; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP84]], ptr [[TMP104]], align 4
790; VF-FOUR-CHECK-NEXT:    [[TMP106:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 20
791; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP85]], ptr [[TMP106]], align 4
792; VF-FOUR-CHECK-NEXT:    [[TMP108:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 24
793; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP86]], ptr [[TMP108]], align 4
794; VF-FOUR-CHECK-NEXT:    [[TMP110:%.*]] = getelementptr inbounds float, ptr [[TMP88]], i32 28
795; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP87]], ptr [[TMP110]], align 4
796; VF-FOUR-CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 32
797; VF-FOUR-CHECK-NEXT:    [[TMP112:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
798; VF-FOUR-CHECK-NEXT:    br i1 [[TMP112]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOPID_MV_CM:![0-9]+]]
799; VF-FOUR-CHECK:       middle.block:
800; VF-FOUR-CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[WIDE_TRIP_COUNT]], [[N_VEC]]
801; VF-FOUR-CHECK-NEXT:    br i1 [[CMP_N]], label [[FOR_END_LOOPEXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
802; VF-FOUR-CHECK:       vec.epilog.iter.check:
803; VF-FOUR-CHECK-NEXT:    [[IND_END19:%.*]] = trunc i64 [[N_VEC]] to i32
804; VF-FOUR-CHECK-NEXT:    [[N_VEC_REMAINING:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_VEC]]
805; VF-FOUR-CHECK-NEXT:    [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_VEC_REMAINING]], 4
806; VF-FOUR-CHECK-NEXT:    br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
807; VF-FOUR-CHECK:       vec.epilog.ph:
808; VF-FOUR-CHECK-NEXT:    [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
809; VF-FOUR-CHECK-NEXT:    [[N_MOD_VF16:%.*]] = urem i64 [[WIDE_TRIP_COUNT]], 4
810; VF-FOUR-CHECK-NEXT:    [[N_VEC17:%.*]] = sub i64 [[WIDE_TRIP_COUNT]], [[N_MOD_VF16]]
811; VF-FOUR-CHECK-NEXT:    [[IND_END:%.*]] = trunc i64 [[N_VEC17]] to i32
812; VF-FOUR-CHECK-NEXT:    br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
813; VF-FOUR-CHECK:       vec.epilog.vector.body:
814; VF-FOUR-CHECK-NEXT:    [[OFFSET_IDX23:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT26:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
815; VF-FOUR-CHECK-NEXT:    [[OFFSET_IDX22:%.*]] = trunc i64 [[OFFSET_IDX23]] to i32
816; VF-FOUR-CHECK-NEXT:    [[TMP113:%.*]] = add i32 [[OFFSET_IDX22]], 0
817; VF-FOUR-CHECK-NEXT:    [[TMP114:%.*]] = add i64 [[OFFSET_IDX23]], 0
818; VF-FOUR-CHECK-NEXT:    [[TMP115:%.*]] = xor i32 [[TMP113]], -1
819; VF-FOUR-CHECK-NEXT:    [[TMP116:%.*]] = add i32 [[TMP115]], [[N]]
820; VF-FOUR-CHECK-NEXT:    [[TMP117:%.*]] = sext i32 [[TMP116]] to i64
821; VF-FOUR-CHECK-NEXT:    [[TMP118:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[TMP117]]
822; VF-FOUR-CHECK-NEXT:    [[TMP119:%.*]] = getelementptr inbounds float, ptr [[TMP118]], i32 0
823; VF-FOUR-CHECK-NEXT:    [[TMP120:%.*]] = getelementptr inbounds float, ptr [[TMP119]], i32 -3
824; VF-FOUR-CHECK-NEXT:    [[WIDE_LOAD24:%.*]] = load <4 x float>, ptr [[TMP120]], align 4
825; VF-FOUR-CHECK-NEXT:    [[REVERSE25:%.*]] = shufflevector <4 x float> [[WIDE_LOAD24]], <4 x float> poison, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
826; VF-FOUR-CHECK-NEXT:    [[TMP122:%.*]] = fadd fast <4 x float> [[REVERSE25]], <float 1.000000e+00, float 1.000000e+00, float 1.000000e+00, float 1.000000e+00>
827; VF-FOUR-CHECK-NEXT:    [[TMP123:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP114]]
828; VF-FOUR-CHECK-NEXT:    [[TMP124:%.*]] = getelementptr inbounds float, ptr [[TMP123]], i32 0
829; VF-FOUR-CHECK-NEXT:    store <4 x float> [[TMP122]], ptr [[TMP124]], align 4
830; VF-FOUR-CHECK-NEXT:    [[INDEX_NEXT26]] = add nuw i64 [[OFFSET_IDX23]], 4
831; VF-FOUR-CHECK-NEXT:    [[TMP126:%.*]] = icmp eq i64 [[INDEX_NEXT26]], [[N_VEC17]]
832; VF-FOUR-CHECK-NEXT:    br i1 [[TMP126]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOPID_EV_CM:![0-9]+]]
833; VF-FOUR-CHECK:       vec.epilog.middle.block:
834; VF-FOUR-CHECK-NEXT:    [[CMP_N20:%.*]] = icmp eq i64 [[WIDE_TRIP_COUNT]], [[N_VEC17]]
835; VF-FOUR-CHECK-NEXT:    br i1 [[CMP_N20]], label [[FOR_END_LOOPEXIT]], label [[VEC_EPILOG_SCALAR_PH]]
836; VF-FOUR-CHECK:       vec.epilog.scalar.ph:
837; VF-FOUR-CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC17]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_SCEVCHECK]] ], [ 0, [[ITER_CHECK]] ]
838; VF-FOUR-CHECK-NEXT:    [[BC_RESUME_VAL18:%.*]] = phi i32 [ [[IND_END]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END19]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_SCEVCHECK]] ], [ 0, [[ITER_CHECK]] ]
839; VF-FOUR-CHECK-NEXT:    br label [[FOR_BODY:%.*]]
840; VF-FOUR-CHECK:       for.body:
841; VF-FOUR-CHECK-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
842; VF-FOUR-CHECK-NEXT:    [[I_014:%.*]] = phi i32 [ [[BC_RESUME_VAL18]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[INC:%.*]], [[FOR_BODY]] ]
843; VF-FOUR-CHECK-NEXT:    [[TMP127:%.*]] = xor i32 [[I_014]], -1
844; VF-FOUR-CHECK-NEXT:    [[SUB2:%.*]] = add i32 [[TMP127]], [[N]]
845; VF-FOUR-CHECK-NEXT:    [[IDXPROM:%.*]] = sext i32 [[SUB2]] to i64
846; VF-FOUR-CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[IDXPROM]]
847; VF-FOUR-CHECK-NEXT:    [[TMP128:%.*]] = load float, ptr [[ARRAYIDX]], align 4
848; VF-FOUR-CHECK-NEXT:    [[CONV3:%.*]] = fadd fast float [[TMP128]], 1.000000e+00
849; VF-FOUR-CHECK-NEXT:    [[ARRAYIDX5:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[INDVARS_IV]]
850; VF-FOUR-CHECK-NEXT:    store float [[CONV3]], ptr [[ARRAYIDX5]], align 4
851; VF-FOUR-CHECK-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
852; VF-FOUR-CHECK-NEXT:    [[INC]] = add nuw nsw i32 [[I_014]], 1
853; VF-FOUR-CHECK-NEXT:    [[EXITCOND:%.*]] = icmp ne i64 [[INDVARS_IV_NEXT]], [[WIDE_TRIP_COUNT]]
854; VF-FOUR-CHECK-NEXT:    br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT]], !llvm.loop [[LOOP7:![0-9]+]]
855; VF-FOUR-CHECK:       for.end.loopexit:
856; VF-FOUR-CHECK-NEXT:    br label [[FOR_END]]
857; VF-FOUR-CHECK:       for.end:
858; VF-FOUR-CHECK-NEXT:    ret i32 0
859;
860entry:
861  %cmp1 = icmp sgt i32 %n, 1
862  br i1 %cmp1, label %for.body.preheader, label %for.end
863
864for.body.preheader:                               ; preds = %entry
865  %0 = add i32 %n, -1
866  %wide.trip.count = zext i32 %0 to i64
867  br label %for.body
868
869for.body:                                         ; preds = %for.body.preheader, %for.body
870  %indvars.iv = phi i64 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
871  %i.014 = phi i32 [ 0, %for.body.preheader ], [ %inc, %for.body ]
872  %1 = xor i32 %i.014, -1
873  %sub2 = add i32 %1, %n
874  %idxprom = sext i32 %sub2 to i64
875  %arrayidx = getelementptr inbounds float, ptr %B, i64 %idxprom
876  %2 = load float, ptr %arrayidx, align 4
877  %conv3 = fadd fast float %2, 1.000000e+00
878  %arrayidx5 = getelementptr inbounds float, ptr %A, i64 %indvars.iv
879  store float %conv3, ptr %arrayidx5, align 4
880  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
881  %inc = add nuw nsw i32 %i.014, 1
882  %exitcond = icmp ne i64 %indvars.iv.next, %wide.trip.count
883  br i1 %exitcond, label %for.body, label %for.end.loopexit
884
885for.end.loopexit:                                 ; preds = %for.body
886  br label %for.end
887
888for.end:                                          ; preds = %for.end.loopexit, %entry
889  ret i32 0
890}
891
892; VF-TWO-CHECK-DAG: [[LOOPID_MV]] = distinct !{[[LOOPID_MV]], [[LOOPID_DISABLE_VECT:!.*]], [[LOOPID_DISABLE_UNROLL:!.*]]}
893; VF-TWO-CHECK-DAG: [[LOOPID_EV]] = distinct !{[[LOOPID_EV]], [[LOOPID_DISABLE_VECT]], [[LOOPID_DISABLE_UNROLL]]}
894; VF-TWO-CHECK-DAG: [[LOOPID_DISABLE_VECT]] = [[DISABLE_VECT_STR:!{!"llvm.loop.isvectorized".*}.*]]
895; VF-TWO-CHECK-DAG: [[LOOPID_DISABLE_UNROLL]] = [[DISABLE_UNROLL_STR:!{!"llvm.loop.unroll.runtime.disable"}.*]]
896;
897; VF-FOUR-CHECK-DAG: [[LOOPID_MV_CM]] = distinct !{[[LOOPID_MV_CM]], [[LOOPID_DISABLE_VECT_CM:!.*]], [[LOOPID_DISABLE_UNROLL_CM:!.*]]}
898; VF-FOUR-CHECK-DAG: [[LOOPID_EV_CM]] = distinct !{[[LOOPID_EV_CM]], [[LOOPID_DISABLE_VECT_CM]], [[LOOPID_DISABLE_UNROLL_CM]]}
899; VF-FOUR-CHECK-DAG: [[LOOPID_DISABLE_VECT_CM]] = [[DISABLE_VECT_STR_CM:!{!"llvm.loop.isvectorized".*}.*]]
900; VF-FOUR-CHECK-DAG: [[LOOPID_DISABLE_UNROLL_CM]] = [[DISABLE_UNROLL_STR_CM:!{!"llvm.loop.unroll.runtime.disable"}.*]]
901;
902attributes #0 = { nounwind "correctly-rounded-divide-sqrt-fp-math"="false" "disable-tail-calls"="false" "frame-pointer"="none" "less-precise-fpmad"="false" "min-legal-vector-width"="0" "no-infs-fp-math"="true" "no-jump-tables"="false" "no-nans-fp-math"="true" "no-signed-zeros-fp-math"="true" "no-trapping-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="ppc64le" "target-features"="+altivec,+bpermd,+crypto,+direct-move,+extdiv,+htm,+power8-vector,+vsx,-power9-vector,-spe" "unsafe-fp-math"="true" "use-soft-float"="false" }
903