1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mcpu=pwr9 -verify-machineinstrs -mtriple=powerpc64le-unknown-linux-gnu < %s | FileCheck %s
; void foo(float *data, float d) {
; long i;
; for (i = 0; i < 8000; i++)
; data[i] = d;
; }
;
; This loop will be unrolled by 96 and vectorized on power9.
; icmp for loop iteration index and loop trip count(384) has LSRUse for 'reg({0,+,384})'.
; Make sure above icmp does not impact LSR choose best formulae sets based on 'reg({(192 + %0),+,384})'
define void @foo(float* nocapture %data, float %d) {
; CHECK-LABEL: foo:
; CHECK: .LBB0_1: # %vector.body
; CHECK: stxv 0, -192(4)
; CHECK-NEXT: stxv 0, -176(4)
; CHECK-NEXT: stxv 0, -160(4)
; CHECK-NEXT: stxv 0, -144(4)
; CHECK-NEXT: stxv 0, -128(4)
; CHECK-NEXT: stxv 0, -112(4)
; CHECK-NEXT: stxv 0, -96(4)
; CHECK-NEXT: stxv 0, -80(4)
; CHECK-NEXT: stxv 0, -64(4)
; CHECK-NEXT: stxv 0, -48(4)
; CHECK-NEXT: stxv 0, -32(4)
; CHECK-NEXT: stxv 0, -16(4)
; CHECK-NEXT: stxv 0, 0(4)
; CHECK-NEXT: stxv 0, 16(4)
; CHECK-NEXT: stxv 0, 32(4)
; CHECK-NEXT: stxv 0, 48(4)
; CHECK-NEXT: stxv 0, 64(4)
; CHECK-NEXT: stxv 0, 80(4)
; CHECK-NEXT: stxv 0, 96(4)
; CHECK-NEXT: stxv 0, 112(4)
; CHECK-NEXT: stxv 0, 128(4)
; CHECK-NEXT: stxv 0, 144(4)
; CHECK-NEXT: stxv 0, 160(4)
; CHECK-NEXT: stxv 0, 176(4)
; CHECK-NEXT: addi 4, 4, 384
; CHECK-NEXT: bdnz .LBB0_1
entry:
%broadcast.splatinsert16 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat17 = shufflevector <4 x float> %broadcast.splatinsert16, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert18 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat19 = shufflevector <4 x float> %broadcast.splatinsert18, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert20 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat21 = shufflevector <4 x float> %broadcast.splatinsert20, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert22 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat23 = shufflevector <4 x float> %broadcast.splatinsert22, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert24 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat25 = shufflevector <4 x float> %broadcast.splatinsert24, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert26 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat27 = shufflevector <4 x float> %broadcast.splatinsert26, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert28 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat29 = shufflevector <4 x float> %broadcast.splatinsert28, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert30 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat31 = shufflevector <4 x float> %broadcast.splatinsert30, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert32 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat33 = shufflevector <4 x float> %broadcast.splatinsert32, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert34 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat35 = shufflevector <4 x float> %broadcast.splatinsert34, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert36 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat37 = shufflevector <4 x float> %broadcast.splatinsert36, <4 x float> undef, <4 x i32> zeroinitializer
%broadcast.splatinsert38 = insertelement <4 x float> undef, float %d, i32 0
%broadcast.splat39 = shufflevector <4 x float> %broadcast.splatinsert38, <4 x float> undef, <4 x i32> zeroinitializer
br label %vector.body
vector.body: ; preds = %vector.body, %entry
%index = phi i64 [ 0, %entry ], [ %index.next.1, %vector.body ]
%0 = getelementptr inbounds float, float* %data, i64 %index
%1 = bitcast float* %0 to <4 x float>*
store <4 x float> %broadcast.splat17, <4 x float>* %1, align 4
%2 = getelementptr inbounds float, float* %0, i64 4
%3 = bitcast float* %2 to <4 x float>*
store <4 x float> %broadcast.splat19, <4 x float>* %3, align 4
%4 = getelementptr inbounds float, float* %0, i64 8
%5 = bitcast float* %4 to <4 x float>*
store <4 x float> %broadcast.splat21, <4 x float>* %5, align 4
%6 = getelementptr inbounds float, float* %0, i64 12
%7 = bitcast float* %6 to <4 x float>*
store <4 x float> %broadcast.splat23, <4 x float>* %7, align 4
%8 = getelementptr inbounds float, float* %0, i64 16
%9 = bitcast float* %8 to <4 x float>*
store <4 x float> %broadcast.splat25, <4 x float>* %9, align 4
%10 = getelementptr inbounds float, float* %0, i64 20
%11 = bitcast float* %10 to <4 x float>*
store <4 x float> %broadcast.splat27, <4 x float>* %11, align 4
%12 = getelementptr inbounds float, float* %0, i64 24
%13 = bitcast float* %12 to <4 x float>*
store <4 x float> %broadcast.splat29, <4 x float>* %13, align 4
%14 = getelementptr inbounds float, float* %0, i64 28
%15 = bitcast float* %14 to <4 x float>*
store <4 x float> %broadcast.splat31, <4 x float>* %15, align 4
%16 = getelementptr inbounds float, float* %0, i64 32
%17 = bitcast float* %16 to <4 x float>*
store <4 x float> %broadcast.splat33, <4 x float>* %17, align 4
%18 = getelementptr inbounds float, float* %0, i64 36
%19 = bitcast float* %18 to <4 x float>*
store <4 x float> %broadcast.splat35, <4 x float>* %19, align 4
%20 = getelementptr inbounds float, float* %0, i64 40
%21 = bitcast float* %20 to <4 x float>*
store <4 x float> %broadcast.splat37, <4 x float>* %21, align 4
%22 = getelementptr inbounds float, float* %0, i64 44
%23 = bitcast float* %22 to <4 x float>*
store <4 x float> %broadcast.splat39, <4 x float>* %23, align 4
%index.next = add nuw nsw i64 %index, 48
%24 = getelementptr inbounds float, float* %data, i64 %index.next
%25 = bitcast float* %24 to <4 x float>*
store <4 x float> %broadcast.splat17, <4 x float>* %25, align 4
%26 = getelementptr inbounds float, float* %24, i64 4
%27 = bitcast float* %26 to <4 x float>*
store <4 x float> %broadcast.splat19, <4 x float>* %27, align 4
%28 = getelementptr inbounds float, float* %24, i64 8
%29 = bitcast float* %28 to <4 x float>*
store <4 x float> %broadcast.splat21, <4 x float>* %29, align 4
%30 = getelementptr inbounds float, float* %24, i64 12
%31 = bitcast float* %30 to <4 x float>*
store <4 x float> %broadcast.splat23, <4 x float>* %31, align 4
%32 = getelementptr inbounds float, float* %24, i64 16
%33 = bitcast float* %32 to <4 x float>*
store <4 x float> %broadcast.splat25, <4 x float>* %33, align 4
%34 = getelementptr inbounds float, float* %24, i64 20
%35 = bitcast float* %34 to <4 x float>*
store <4 x float> %broadcast.splat27, <4 x float>* %35, align 4
%36 = getelementptr inbounds float, float* %24, i64 24
%37 = bitcast float* %36 to <4 x float>*
store <4 x float> %broadcast.splat29, <4 x float>* %37, align 4
%38 = getelementptr inbounds float, float* %24, i64 28
%39 = bitcast float* %38 to <4 x float>*
store <4 x float> %broadcast.splat31, <4 x float>* %39, align 4
%40 = getelementptr inbounds float, float* %24, i64 32
%41 = bitcast float* %40 to <4 x float>*
store <4 x float> %broadcast.splat33, <4 x float>* %41, align 4
%42 = getelementptr inbounds float, float* %24, i64 36
%43 = bitcast float* %42 to <4 x float>*
store <4 x float> %broadcast.splat35, <4 x float>* %43, align 4
%44 = getelementptr inbounds float, float* %24, i64 40
%45 = bitcast float* %44 to <4 x float>*
store <4 x float> %broadcast.splat37, <4 x float>* %45, align 4
%46 = getelementptr inbounds float, float* %24, i64 44
%47 = bitcast float* %46 to <4 x float>*
store <4 x float> %broadcast.splat39, <4 x float>* %47, align 4
%index.next.1 = add nuw nsw i64 %index, 96
%48 = icmp eq i64 %index.next.1, 7968
br i1 %48, label %for.body, label %vector.body
for.body: ; preds = %vector.body
%arrayidx = getelementptr inbounds float, float* %data, i64 7968
store float %d, float* %arrayidx, align 4
%arrayidx.1 = getelementptr inbounds float, float* %data, i64 7969
store float %d, float* %arrayidx.1, align 4
%arrayidx.2 = getelementptr inbounds float, float* %data, i64 7970
store float %d, float* %arrayidx.2, align 4
%arrayidx.3 = getelementptr inbounds float, float* %data, i64 7971
store float %d, float* %arrayidx.3, align 4
%arrayidx.4 = getelementptr inbounds float, float* %data, i64 7972
store float %d, float* %arrayidx.4, align 4
%arrayidx.5 = getelementptr inbounds float, float* %data, i64 7973
store float %d, float* %arrayidx.5, align 4
%arrayidx.6 = getelementptr inbounds float, float* %data, i64 7974
store float %d, float* %arrayidx.6, align 4
%arrayidx.7 = getelementptr inbounds float, float* %data, i64 7975
store float %d, float* %arrayidx.7, align 4
%arrayidx.8 = getelementptr inbounds float, float* %data, i64 7976
store float %d, float* %arrayidx.8, align 4
%arrayidx.9 = getelementptr inbounds float, float* %data, i64 7977
store float %d, float* %arrayidx.9, align 4
%arrayidx.10 = getelementptr inbounds float, float* %data, i64 7978
store float %d, float* %arrayidx.10, align 4
%arrayidx.11 = getelementptr inbounds float, float* %data, i64 7979
store float %d, float* %arrayidx.11, align 4
%arrayidx.12 = getelementptr inbounds float, float* %data, i64 7980
store float %d, float* %arrayidx.12, align 4
%arrayidx.13 = getelementptr inbounds float, float* %data, i64 7981
store float %d, float* %arrayidx.13, align 4
%arrayidx.14 = getelementptr inbounds float, float* %data, i64 7982
store float %d, float* %arrayidx.14, align 4
%arrayidx.15 = getelementptr inbounds float, float* %data, i64 7983
store float %d, float* %arrayidx.15, align 4
%arrayidx.16 = getelementptr inbounds float, float* %data, i64 7984
store float %d, float* %arrayidx.16, align 4
%arrayidx.17 = getelementptr inbounds float, float* %data, i64 7985
store float %d, float* %arrayidx.17, align 4
%arrayidx.18 = getelementptr inbounds float, float* %data, i64 7986
store float %d, float* %arrayidx.18, align 4
%arrayidx.19 = getelementptr inbounds float, float* %data, i64 7987
store float %d, float* %arrayidx.19, align 4
%arrayidx.20 = getelementptr inbounds float, float* %data, i64 7988
store float %d, float* %arrayidx.20, align 4
%arrayidx.21 = getelementptr inbounds float, float* %data, i64 7989
store float %d, float* %arrayidx.21, align 4
%arrayidx.22 = getelementptr inbounds float, float* %data, i64 7990
store float %d, float* %arrayidx.22, align 4
%arrayidx.23 = getelementptr inbounds float, float* %data, i64 7991
store float %d, float* %arrayidx.23, align 4
%arrayidx.24 = getelementptr inbounds float, float* %data, i64 7992
store float %d, float* %arrayidx.24, align 4
%arrayidx.25 = getelementptr inbounds float, float* %data, i64 7993
store float %d, float* %arrayidx.25, align 4
%arrayidx.26 = getelementptr inbounds float, float* %data, i64 7994
store float %d, float* %arrayidx.26, align 4
%arrayidx.27 = getelementptr inbounds float, float* %data, i64 7995
store float %d, float* %arrayidx.27, align 4
%arrayidx.28 = getelementptr inbounds float, float* %data, i64 7996
store float %d, float* %arrayidx.28, align 4
%arrayidx.29 = getelementptr inbounds float, float* %data, i64 7997
store float %d, float* %arrayidx.29, align 4
%arrayidx.30 = getelementptr inbounds float, float* %data, i64 7998
store float %d, float* %arrayidx.30, align 4
%arrayidx.31 = getelementptr inbounds float, float* %data, i64 7999
store float %d, float* %arrayidx.31, align 4
ret void
}
|