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
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=X32
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=X64
define <4 x double> @test_broadcast_2f64_4f64(<2 x double> *%p) nounwind {
; X32-LABEL: test_broadcast_2f64_4f64:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_2f64_4f64:
; X64: # %bb.0:
; X64-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X64-NEXT: retq
%1 = load <2 x double>, <2 x double> *%p
%2 = shufflevector <2 x double> %1, <2 x double> undef, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
ret <4 x double> %2
}
define <4 x i64> @test_broadcast_2i64_4i64(<2 x i64> *%p) nounwind {
; X32-LABEL: test_broadcast_2i64_4i64:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_2i64_4i64:
; X64: # %bb.0:
; X64-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X64-NEXT: retq
%1 = load <2 x i64>, <2 x i64> *%p
%2 = shufflevector <2 x i64> %1, <2 x i64> undef, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
ret <4 x i64> %2
}
define <8 x float> @test_broadcast_4f32_8f32(<4 x float> *%p) nounwind {
; X32-LABEL: test_broadcast_4f32_8f32:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_4f32_8f32:
; X64: # %bb.0:
; X64-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X64-NEXT: retq
%1 = load <4 x float>, <4 x float> *%p
%2 = shufflevector <4 x float> %1, <4 x float> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
ret <8 x float> %2
}
define <8 x i32> @test_broadcast_4i32_8i32(<4 x i32> *%p) nounwind {
; X32-LABEL: test_broadcast_4i32_8i32:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_4i32_8i32:
; X64: # %bb.0:
; X64-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X64-NEXT: retq
%1 = load <4 x i32>, <4 x i32> *%p
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
ret <8 x i32> %2
}
define <16 x i16> @test_broadcast_8i16_16i16(<8 x i16> *%p) nounwind {
; X32-LABEL: test_broadcast_8i16_16i16:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_8i16_16i16:
; X64: # %bb.0:
; X64-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X64-NEXT: retq
%1 = load <8 x i16>, <8 x i16> *%p
%2 = shufflevector <8 x i16> %1, <8 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
ret <16 x i16> %2
}
define <32 x i8> @test_broadcast_16i8_32i8(<16 x i8> *%p) nounwind {
; X32-LABEL: test_broadcast_16i8_32i8:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_16i8_32i8:
; X64: # %bb.0:
; X64-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X64-NEXT: retq
%1 = load <16 x i8>, <16 x i8> *%p
%2 = shufflevector <16 x i8> %1, <16 x i8> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
ret <32 x i8> %2
}
; PR38949 - https://bugs.llvm.org/show_bug.cgi?id=38949
; Don't limit the transform based on extra uses of the load itself (the store is a user of the load's chain value).
define void @subv_reuse_is_ok(<4 x float>* %a, <8 x float>* %b) {
; X32-LABEL: subv_reuse_is_ok:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X32-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X32-NEXT: vmovups %ymm0, (%eax)
; X32-NEXT: vzeroupper
; X32-NEXT: retl
;
; X64-LABEL: subv_reuse_is_ok:
; X64: # %bb.0:
; X64-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X64-NEXT: vmovups %ymm0, (%rsi)
; X64-NEXT: vzeroupper
; X64-NEXT: retq
%ld = load <4 x float>, <4 x float>* %a, align 1
%splat128 = shufflevector <4 x float> %ld, <4 x float> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
store <8 x float> %splat128, <8 x float>* %b, align 16
ret void
}
define <4 x double> @test_broadcast_2f64_4f64_reuse(<2 x double>* %p0, <2 x double>* %p1) {
; X32-LABEL: test_broadcast_2f64_4f64_reuse:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X32-NEXT: vmovaps (%ecx), %xmm1
; X32-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X32-NEXT: vmovaps %xmm1, (%eax)
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_2f64_4f64_reuse:
; X64: # %bb.0:
; X64-NEXT: vmovaps (%rdi), %xmm1
; X64-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X64-NEXT: vmovaps %xmm1, (%rsi)
; X64-NEXT: retq
%1 = load <2 x double>, <2 x double>* %p0
%2 = shufflevector <2 x double> %1, <2 x double> undef, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
store <2 x double> %1, <2 x double>* %p1
ret <4 x double> %2
}
define <4 x i64> @test_broadcast_2i64_4i64_reuse(<2 x i64>* %p0, <2 x i64>* %p1) {
; X32-LABEL: test_broadcast_2i64_4i64_reuse:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X32-NEXT: vmovaps (%ecx), %xmm1
; X32-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X32-NEXT: vmovaps %xmm1, (%eax)
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_2i64_4i64_reuse:
; X64: # %bb.0:
; X64-NEXT: vmovaps (%rdi), %xmm1
; X64-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X64-NEXT: vmovaps %xmm1, (%rsi)
; X64-NEXT: retq
%1 = load <2 x i64>, <2 x i64>* %p0
%2 = shufflevector <2 x i64> %1, <2 x i64> undef, <4 x i32> <i32 0, i32 1, i32 0, i32 1>
store <2 x i64> %1, <2 x i64>* %p1
ret <4 x i64> %2
}
define <8 x float> @test_broadcast_4f32_8f32_reuse(<4 x float>* %p0, <4 x float>* %p1) {
; X32-LABEL: test_broadcast_4f32_8f32_reuse:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X32-NEXT: vmovaps (%ecx), %xmm1
; X32-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X32-NEXT: vmovaps %xmm1, (%eax)
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_4f32_8f32_reuse:
; X64: # %bb.0:
; X64-NEXT: vmovaps (%rdi), %xmm1
; X64-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X64-NEXT: vmovaps %xmm1, (%rsi)
; X64-NEXT: retq
%1 = load <4 x float>, <4 x float>* %p0
%2 = shufflevector <4 x float> %1, <4 x float> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
store <4 x float> %1, <4 x float>* %p1
ret <8 x float> %2
}
define <8 x i32> @test_broadcast_4i32_8i32_reuse(<4 x i32>* %p0, <4 x i32>* %p1) {
; X32-LABEL: test_broadcast_4i32_8i32_reuse:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X32-NEXT: vmovaps (%ecx), %xmm1
; X32-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X32-NEXT: vmovaps %xmm1, (%eax)
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_4i32_8i32_reuse:
; X64: # %bb.0:
; X64-NEXT: vmovaps (%rdi), %xmm1
; X64-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X64-NEXT: vmovaps %xmm1, (%rsi)
; X64-NEXT: retq
%1 = load <4 x i32>, <4 x i32>* %p0
%2 = shufflevector <4 x i32> %1, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
store <4 x i32> %1, <4 x i32>* %p1
ret <8 x i32> %2
}
define <16 x i16> @test_broadcast_8i16_16i16_reuse(<8 x i16> *%p0, <8 x i16> *%p1) nounwind {
; X32-LABEL: test_broadcast_8i16_16i16_reuse:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X32-NEXT: vmovaps (%ecx), %xmm1
; X32-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X32-NEXT: vmovaps %xmm1, (%eax)
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_8i16_16i16_reuse:
; X64: # %bb.0:
; X64-NEXT: vmovaps (%rdi), %xmm1
; X64-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X64-NEXT: vmovaps %xmm1, (%rsi)
; X64-NEXT: retq
%1 = load <8 x i16>, <8 x i16> *%p0
%2 = shufflevector <8 x i16> %1, <8 x i16> undef, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
store <8 x i16> %1, <8 x i16>* %p1
ret <16 x i16> %2
}
define <32 x i8> @test_broadcast_16i8_32i8_reuse(<16 x i8> *%p0, <16 x i8> *%p1) nounwind {
; X32-LABEL: test_broadcast_16i8_32i8_reuse:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X32-NEXT: vmovaps (%ecx), %xmm1
; X32-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X32-NEXT: vmovaps %xmm1, (%eax)
; X32-NEXT: retl
;
; X64-LABEL: test_broadcast_16i8_32i8_reuse:
; X64: # %bb.0:
; X64-NEXT: vmovaps (%rdi), %xmm1
; X64-NEXT: vinsertf128 $1, %xmm1, %ymm1, %ymm0
; X64-NEXT: vmovaps %xmm1, (%rsi)
; X64-NEXT: retq
%1 = load <16 x i8>, <16 x i8> *%p0
%2 = shufflevector <16 x i8> %1, <16 x i8> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
store <16 x i8> %1, <16 x i8>* %p1
ret <32 x i8> %2
}
define <8 x i32> @PR29088(<4 x i32>* %p0, <8 x float>* %p1) {
; X32-LABEL: PR29088:
; X32: # %bb.0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X32-NEXT: vxorps %xmm1, %xmm1, %xmm1
; X32-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X32-NEXT: vmovaps %ymm1, (%eax)
; X32-NEXT: retl
;
; X64-LABEL: PR29088:
; X64: # %bb.0:
; X64-NEXT: vxorps %xmm1, %xmm1, %xmm1
; X64-NEXT: vbroadcastf128 {{.*#+}} ymm0 = mem[0,1,0,1]
; X64-NEXT: vmovaps %ymm1, (%rsi)
; X64-NEXT: retq
%ld = load <4 x i32>, <4 x i32>* %p0
store <8 x float> zeroinitializer, <8 x float>* %p1
%shuf = shufflevector <4 x i32> %ld, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
ret <8 x i32> %shuf
}
|