1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2; RUN: opt -slp-vectorizer -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=skylake -slp-threshold=-7 | FileCheck %s --check-prefix=CHECK
3; RUN: opt -slp-vectorizer -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=skylake -slp-threshold=-8 -slp-min-tree-size=6 | FileCheck %s --check-prefix=FORCE_REDUCTION
4
5define void @Test(i32) {
6; CHECK-LABEL: @Test(
7; CHECK-NEXT:  entry:
8; CHECK-NEXT:    br label [[LOOP:%.*]]
9; CHECK:       loop:
10; CHECK-NEXT:    [[TMP1:%.*]] = phi <2 x i32> [ [[TMP15:%.*]], [[LOOP]] ], [ zeroinitializer, [[ENTRY:%.*]] ]
11; CHECK-NEXT:    [[SHUFFLE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> undef, <8 x i32> <i32 0, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1, i32 1>
12; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <8 x i32> [[SHUFFLE]], i32 1
13; CHECK-NEXT:    [[TMP3:%.*]] = add <8 x i32> [[SHUFFLE]], <i32 0, i32 55, i32 285, i32 1240, i32 1496, i32 8555, i32 12529, i32 13685>
14; CHECK-NEXT:    [[TMP4:%.*]] = call i32 @llvm.vector.reduce.and.v8i32(<8 x i32> [[TMP3]])
15; CHECK-NEXT:    [[OP_EXTRA:%.*]] = and i32 [[TMP4]], [[TMP0:%.*]]
16; CHECK-NEXT:    [[OP_EXTRA1:%.*]] = and i32 [[OP_EXTRA]], [[TMP0]]
17; CHECK-NEXT:    [[OP_EXTRA2:%.*]] = and i32 [[OP_EXTRA1]], [[TMP0]]
18; CHECK-NEXT:    [[OP_EXTRA3:%.*]] = and i32 [[OP_EXTRA2]], [[TMP0]]
19; CHECK-NEXT:    [[OP_EXTRA4:%.*]] = and i32 [[OP_EXTRA3]], [[TMP0]]
20; CHECK-NEXT:    [[OP_EXTRA5:%.*]] = and i32 [[OP_EXTRA4]], [[TMP0]]
21; CHECK-NEXT:    [[OP_EXTRA6:%.*]] = and i32 [[OP_EXTRA5]], [[TMP0]]
22; CHECK-NEXT:    [[OP_EXTRA7:%.*]] = and i32 [[OP_EXTRA6]], [[TMP0]]
23; CHECK-NEXT:    [[OP_EXTRA8:%.*]] = and i32 [[OP_EXTRA7]], [[TMP0]]
24; CHECK-NEXT:    [[OP_EXTRA9:%.*]] = and i32 [[OP_EXTRA8]], [[TMP0]]
25; CHECK-NEXT:    [[OP_EXTRA10:%.*]] = and i32 [[OP_EXTRA9]], [[TMP0]]
26; CHECK-NEXT:    [[OP_EXTRA11:%.*]] = and i32 [[OP_EXTRA10]], [[TMP0]]
27; CHECK-NEXT:    [[OP_EXTRA12:%.*]] = and i32 [[OP_EXTRA11]], [[TMP0]]
28; CHECK-NEXT:    [[OP_EXTRA13:%.*]] = and i32 [[OP_EXTRA12]], [[TMP0]]
29; CHECK-NEXT:    [[OP_EXTRA14:%.*]] = and i32 [[OP_EXTRA13]], [[TMP0]]
30; CHECK-NEXT:    [[OP_EXTRA15:%.*]] = and i32 [[OP_EXTRA14]], [[TMP0]]
31; CHECK-NEXT:    [[OP_EXTRA16:%.*]] = and i32 [[OP_EXTRA15]], [[TMP0]]
32; CHECK-NEXT:    [[OP_EXTRA17:%.*]] = and i32 [[OP_EXTRA16]], [[TMP0]]
33; CHECK-NEXT:    [[OP_EXTRA18:%.*]] = and i32 [[OP_EXTRA17]], [[TMP0]]
34; CHECK-NEXT:    [[OP_EXTRA19:%.*]] = and i32 [[OP_EXTRA18]], [[TMP0]]
35; CHECK-NEXT:    [[OP_EXTRA20:%.*]] = and i32 [[OP_EXTRA19]], [[TMP0]]
36; CHECK-NEXT:    [[OP_EXTRA21:%.*]] = and i32 [[OP_EXTRA20]], [[TMP0]]
37; CHECK-NEXT:    [[OP_EXTRA22:%.*]] = and i32 [[OP_EXTRA21]], [[TMP0]]
38; CHECK-NEXT:    [[OP_EXTRA23:%.*]] = and i32 [[OP_EXTRA22]], [[TMP0]]
39; CHECK-NEXT:    [[OP_EXTRA24:%.*]] = and i32 [[OP_EXTRA23]], [[TMP0]]
40; CHECK-NEXT:    [[OP_EXTRA25:%.*]] = and i32 [[OP_EXTRA24]], [[TMP0]]
41; CHECK-NEXT:    [[OP_EXTRA26:%.*]] = and i32 [[OP_EXTRA25]], [[TMP0]]
42; CHECK-NEXT:    [[TMP5:%.*]] = insertelement <2 x i32> undef, i32 [[OP_EXTRA26]], i32 0
43; CHECK-NEXT:    [[TMP6:%.*]] = insertelement <2 x i32> [[TMP5]], i32 14910, i32 1
44; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <2 x i32> undef, i32 [[TMP2]], i32 0
45; CHECK-NEXT:    [[TMP8:%.*]] = insertelement <2 x i32> [[TMP7]], i32 [[TMP2]], i32 1
46; CHECK-NEXT:    [[TMP9:%.*]] = and <2 x i32> [[TMP6]], [[TMP8]]
47; CHECK-NEXT:    [[TMP10:%.*]] = add <2 x i32> [[TMP6]], [[TMP8]]
48; CHECK-NEXT:    [[TMP11:%.*]] = shufflevector <2 x i32> [[TMP9]], <2 x i32> [[TMP10]], <2 x i32> <i32 0, i32 3>
49; CHECK-NEXT:    [[TMP12:%.*]] = extractelement <2 x i32> [[TMP11]], i32 0
50; CHECK-NEXT:    [[TMP13:%.*]] = insertelement <2 x i32> undef, i32 [[TMP12]], i32 0
51; CHECK-NEXT:    [[TMP14:%.*]] = extractelement <2 x i32> [[TMP11]], i32 1
52; CHECK-NEXT:    [[TMP15]] = insertelement <2 x i32> [[TMP13]], i32 [[TMP14]], i32 1
53; CHECK-NEXT:    br label [[LOOP]]
54;
55; FORCE_REDUCTION-LABEL: @Test(
56; FORCE_REDUCTION-NEXT:  entry:
57; FORCE_REDUCTION-NEXT:    br label [[LOOP:%.*]]
58; FORCE_REDUCTION:       loop:
59; FORCE_REDUCTION-NEXT:    [[TMP1:%.*]] = phi <2 x i32> [ [[TMP13:%.*]], [[LOOP]] ], [ zeroinitializer, [[ENTRY:%.*]] ]
60; FORCE_REDUCTION-NEXT:    [[SHUFFLE:%.*]] = shufflevector <2 x i32> [[TMP1]], <2 x i32> undef, <4 x i32> <i32 0, i32 1, i32 1, i32 1>
61; FORCE_REDUCTION-NEXT:    [[TMP2:%.*]] = extractelement <4 x i32> [[SHUFFLE]], i32 1
62; FORCE_REDUCTION-NEXT:    [[TMP3:%.*]] = add <4 x i32> [[SHUFFLE]], <i32 0, i32 55, i32 285, i32 1240>
63; FORCE_REDUCTION-NEXT:    [[VAL_20:%.*]] = add i32 [[TMP2]], 1496
64; FORCE_REDUCTION-NEXT:    [[VAL_34:%.*]] = add i32 [[TMP2]], 8555
65; FORCE_REDUCTION-NEXT:    [[TMP4:%.*]] = call i32 @llvm.vector.reduce.and.v4i32(<4 x i32> [[TMP3]])
66; FORCE_REDUCTION-NEXT:    [[TMP5:%.*]] = and i32 [[TMP4]], [[VAL_20]]
67; FORCE_REDUCTION-NEXT:    [[TMP6:%.*]] = and i32 [[TMP5]], [[VAL_34]]
68; FORCE_REDUCTION-NEXT:    [[OP_EXTRA:%.*]] = and i32 [[TMP6]], [[TMP0:%.*]]
69; FORCE_REDUCTION-NEXT:    [[OP_EXTRA1:%.*]] = and i32 [[OP_EXTRA]], [[TMP0]]
70; FORCE_REDUCTION-NEXT:    [[OP_EXTRA2:%.*]] = and i32 [[OP_EXTRA1]], [[TMP0]]
71; FORCE_REDUCTION-NEXT:    [[OP_EXTRA3:%.*]] = and i32 [[OP_EXTRA2]], [[TMP0]]
72; FORCE_REDUCTION-NEXT:    [[OP_EXTRA4:%.*]] = and i32 [[OP_EXTRA3]], [[TMP0]]
73; FORCE_REDUCTION-NEXT:    [[OP_EXTRA5:%.*]] = and i32 [[OP_EXTRA4]], [[TMP0]]
74; FORCE_REDUCTION-NEXT:    [[OP_EXTRA6:%.*]] = and i32 [[OP_EXTRA5]], [[TMP0]]
75; FORCE_REDUCTION-NEXT:    [[OP_EXTRA7:%.*]] = and i32 [[OP_EXTRA6]], [[TMP0]]
76; FORCE_REDUCTION-NEXT:    [[OP_EXTRA8:%.*]] = and i32 [[OP_EXTRA7]], [[TMP0]]
77; FORCE_REDUCTION-NEXT:    [[OP_EXTRA9:%.*]] = and i32 [[OP_EXTRA8]], [[TMP0]]
78; FORCE_REDUCTION-NEXT:    [[OP_EXTRA10:%.*]] = and i32 [[OP_EXTRA9]], [[TMP0]]
79; FORCE_REDUCTION-NEXT:    [[OP_EXTRA11:%.*]] = and i32 [[OP_EXTRA10]], [[TMP0]]
80; FORCE_REDUCTION-NEXT:    [[OP_EXTRA12:%.*]] = and i32 [[OP_EXTRA11]], [[TMP0]]
81; FORCE_REDUCTION-NEXT:    [[OP_EXTRA13:%.*]] = and i32 [[OP_EXTRA12]], [[TMP0]]
82; FORCE_REDUCTION-NEXT:    [[OP_EXTRA14:%.*]] = and i32 [[OP_EXTRA13]], [[TMP0]]
83; FORCE_REDUCTION-NEXT:    [[OP_EXTRA15:%.*]] = and i32 [[OP_EXTRA14]], [[TMP0]]
84; FORCE_REDUCTION-NEXT:    [[OP_EXTRA16:%.*]] = and i32 [[OP_EXTRA15]], [[TMP0]]
85; FORCE_REDUCTION-NEXT:    [[OP_EXTRA17:%.*]] = and i32 [[OP_EXTRA16]], [[TMP0]]
86; FORCE_REDUCTION-NEXT:    [[OP_EXTRA18:%.*]] = and i32 [[OP_EXTRA17]], [[TMP0]]
87; FORCE_REDUCTION-NEXT:    [[OP_EXTRA19:%.*]] = and i32 [[OP_EXTRA18]], [[TMP0]]
88; FORCE_REDUCTION-NEXT:    [[OP_EXTRA20:%.*]] = and i32 [[OP_EXTRA19]], [[TMP0]]
89; FORCE_REDUCTION-NEXT:    [[OP_EXTRA21:%.*]] = and i32 [[OP_EXTRA20]], [[TMP0]]
90; FORCE_REDUCTION-NEXT:    [[OP_EXTRA22:%.*]] = and i32 [[OP_EXTRA21]], [[TMP0]]
91; FORCE_REDUCTION-NEXT:    [[OP_EXTRA23:%.*]] = and i32 [[OP_EXTRA22]], [[TMP0]]
92; FORCE_REDUCTION-NEXT:    [[OP_EXTRA24:%.*]] = and i32 [[OP_EXTRA23]], [[TMP0]]
93; FORCE_REDUCTION-NEXT:    [[OP_EXTRA25:%.*]] = and i32 [[OP_EXTRA24]], [[TMP0]]
94; FORCE_REDUCTION-NEXT:    [[OP_EXTRA26:%.*]] = and i32 [[OP_EXTRA25]], [[TMP0]]
95; FORCE_REDUCTION-NEXT:    [[OP_EXTRA27:%.*]] = and i32 [[OP_EXTRA26]], [[TMP2]]
96; FORCE_REDUCTION-NEXT:    [[VAL_39:%.*]] = add i32 [[TMP2]], 12529
97; FORCE_REDUCTION-NEXT:    [[VAL_40:%.*]] = and i32 [[OP_EXTRA27]], [[VAL_39]]
98; FORCE_REDUCTION-NEXT:    [[VAL_41:%.*]] = add i32 [[TMP2]], 13685
99; FORCE_REDUCTION-NEXT:    [[TMP7:%.*]] = insertelement <2 x i32> undef, i32 [[VAL_40]], i32 0
100; FORCE_REDUCTION-NEXT:    [[TMP8:%.*]] = insertelement <2 x i32> [[TMP7]], i32 [[TMP2]], i32 1
101; FORCE_REDUCTION-NEXT:    [[TMP9:%.*]] = insertelement <2 x i32> undef, i32 [[VAL_41]], i32 0
102; FORCE_REDUCTION-NEXT:    [[TMP10:%.*]] = insertelement <2 x i32> [[TMP9]], i32 14910, i32 1
103; FORCE_REDUCTION-NEXT:    [[TMP11:%.*]] = and <2 x i32> [[TMP8]], [[TMP10]]
104; FORCE_REDUCTION-NEXT:    [[TMP12:%.*]] = add <2 x i32> [[TMP8]], [[TMP10]]
105; FORCE_REDUCTION-NEXT:    [[TMP13]] = shufflevector <2 x i32> [[TMP11]], <2 x i32> [[TMP12]], <2 x i32> <i32 0, i32 3>
106; FORCE_REDUCTION-NEXT:    br label [[LOOP]]
107;
108entry:
109  br label %loop
110
111loop:
112  %local_4_39.us = phi i32 [ %val_42, %loop ], [ 0, %entry ]
113  %local_8_43.us = phi i32 [ %val_43, %loop ], [ 0, %entry ]
114  %val_0 = add i32 %local_4_39.us, 0
115  %val_1 = and i32 %local_8_43.us, %val_0
116  %val_2 = and i32 %val_1, %0
117  %val_3 = and i32 %val_2, %0
118  %val_4 = and i32 %val_3, %0
119  %val_5 = and i32 %val_4, %0
120  %val_6 = add i32 %local_8_43.us, 55
121  %val_7 = and i32 %val_5, %val_6
122  %val_8 = and i32 %val_7, %0
123  %val_9 = and i32 %val_8, %0
124  %val_10 = and i32 %val_9, %0
125  %val_11 = add i32 %local_8_43.us, 285
126  %val_12 = and i32 %val_10, %val_11
127  %val_13 = and i32 %val_12, %0
128  %val_14 = and i32 %val_13, %0
129  %val_15 = and i32 %val_14, %0
130  %val_16 = and i32 %val_15, %0
131  %val_17 = and i32 %val_16, %0
132  %val_18 = add i32 %local_8_43.us, 1240
133  %val_19 = and i32 %val_17, %val_18
134  %val_20 = add i32 %local_8_43.us, 1496
135  %val_21 = and i32 %val_19, %val_20
136  %val_22 = and i32 %val_21, %0
137  %val_23 = and i32 %val_22, %0
138  %val_24 = and i32 %val_23, %0
139  %val_25 = and i32 %val_24, %0
140  %val_26 = and i32 %val_25, %0
141  %val_27 = and i32 %val_26, %0
142  %val_28 = and i32 %val_27, %0
143  %val_29 = and i32 %val_28, %0
144  %val_30 = and i32 %val_29, %0
145  %val_31 = and i32 %val_30, %0
146  %val_32 = and i32 %val_31, %0
147  %val_33 = and i32 %val_32, %0
148  %val_34 = add i32 %local_8_43.us, 8555
149  %val_35 = and i32 %val_33, %val_34
150  %val_36 = and i32 %val_35, %0
151  %val_37 = and i32 %val_36, %0
152  %val_38 = and i32 %val_37, %0
153  %val_39 = add i32 %local_8_43.us, 12529
154  %val_40 = and i32 %val_38, %val_39
155  %val_41 = add i32 %local_8_43.us, 13685
156  %val_42 = and i32 %val_40, %val_41
157  %val_43 = add i32 %local_8_43.us, 14910
158  br label %loop
159}
160