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
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
|
; RUN: llc %s -mtriple=aarch64 -mattr=+v8.3a,+fullfp16 -o - | FileCheck %s
define <4 x half> @test_16x4(<4 x half> %a, <4 x half> %b, <4 x half> %c) {
entry:
; CHECK-LABEL: test_16x4
; CHECK: fcmla v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, #0
;
%res = tail call <4 x half> @llvm.aarch64.neon.vcmla.rot0.v4f16(<4 x half> %a, <4 x half> %b, <4 x half> %c)
ret <4 x half> %res
}
define <4 x half> @test_16x4_lane_1(<4 x half> %a, <4 x half> %b, <4 x half> %c) {
entry:
; CHECK-LABEL: test_16x4_lane_1
; CHECK: fcmla v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, v{{[0-9]+}}.h[1], #0
;
%c.cast = bitcast <4 x half> %c to <2 x i32>
%c.dup = shufflevector <2 x i32> %c.cast , <2 x i32> undef, <2 x i32> <i32 1, i32 1>
%c.res = bitcast <2 x i32> %c.dup to <4 x half>
%res = tail call <4 x half> @llvm.aarch64.neon.vcmla.rot0.v4f16(<4 x half> %a, <4 x half> %b, <4 x half> %c.res)
ret <4 x half> %res
}
define <4 x half> @test_rot90_16x4(<4 x half> %a, <4 x half> %b, <4 x half> %c) {
entry:
; CHECK-LABEL: test_rot90_16x4
; CHECK: fcmla v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, #90
;
%res = tail call <4 x half> @llvm.aarch64.neon.vcmla.rot90.v4f16(<4 x half> %a, <4 x half> %b, <4 x half> %c)
ret <4 x half> %res
}
define <4 x half> @test_rot90_16x4_lane_0(<4 x half> %a, <4 x half> %b, <4 x half> %c) {
entry:
; CHECK-LABEL: test_rot90_16x4_lane_0
; CHECK: fcmla v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, v{{[0-9]+}}.h[0], #90
;
%c.cast = bitcast <4 x half> %c to <2 x i32>
%c.dup = shufflevector <2 x i32> %c.cast , <2 x i32> undef, <2 x i32> <i32 0, i32 0>
%c.res = bitcast <2 x i32> %c.dup to <4 x half>
%res = tail call <4 x half> @llvm.aarch64.neon.vcmla.rot90.v4f16(<4 x half> %a, <4 x half> %b, <4 x half> %c.res)
ret <4 x half> %res
}
define <4 x half> @test_rot180_16x4(<4 x half> %a, <4 x half> %b, <4 x half> %c) {
entry:
; CHECK-LABEL: test_rot180_16x4
; CHECK: fcmla v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, #180
;
%res = tail call <4 x half> @llvm.aarch64.neon.vcmla.rot180.v4f16(<4 x half> %a, <4 x half> %b, <4 x half> %c)
ret <4 x half> %res
}
define <4 x half> @test_rot180_16x4_lane_0(<4 x half> %a, <4 x half> %b, <8 x half> %c) {
entry:
; CHECK-LABEL: test_rot180_16x4_lane_0
; CHECK: fcmla v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, v{{[0-9]+}}.h[0], #180
%c.cast = bitcast <8 x half> %c to <4 x i32>
%c.dup = shufflevector <4 x i32> %c.cast , <4 x i32> undef, <2 x i32> <i32 0, i32 0>
%c.res = bitcast <2 x i32> %c.dup to <4 x half>
%res = tail call <4 x half> @llvm.aarch64.neon.vcmla.rot180.v4f16(<4 x half> %a, <4 x half> %b, <4 x half> %c.res)
ret <4 x half> %res
}
define <4 x half> @test_rot270_16x4(<4 x half> %a, <4 x half> %b, <4 x half> %c) {
entry:
; CHECK-LABEL: test_rot270_16x4
; CHECK: fcmla v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, v{{[0-9]+}}.4h, #270
;
%res = tail call <4 x half> @llvm.aarch64.neon.vcmla.rot270.v4f16(<4 x half> %a, <4 x half> %b, <4 x half> %c)
ret <4 x half> %res
}
define <2 x float> @test_32x2(<2 x float> %a, <2 x float> %b, <2 x float> %c) {
entry:
; CHECK-LABEL: test_32x2
; CHECK: fcmla v{{[0-9]+}}.2s, v{{[0-9]+}}.2s, v{{[0-9]+}}.2s, #0
;
%res = tail call <2 x float> @llvm.aarch64.neon.vcmla.rot0.v2f32(<2 x float> %a, <2 x float> %b, <2 x float> %c)
ret <2 x float> %res
}
define <2 x float> @test_rot90_32x2(<2 x float> %a, <2 x float> %b, <2 x float> %c) {
entry:
; CHECK-LABEL: test_rot90_32x2
; CHECK: fcmla v{{[0-9]+}}.2s, v{{[0-9]+}}.2s, v{{[0-9]+}}.2s, #90
;
%res = tail call <2 x float> @llvm.aarch64.neon.vcmla.rot90.v2f32(<2 x float> %a, <2 x float> %b, <2 x float> %c)
ret <2 x float> %res
}
define <2 x float> @test_rot180_32x2(<2 x float> %a, <2 x float> %b, <2 x float> %c) {
entry:
; CHECK-LABEL: test_rot180_32x2
; CHECK: fcmla v{{[0-9]+}}.2s, v{{[0-9]+}}.2s, v{{[0-9]+}}.2s, #180
;
%res = tail call <2 x float> @llvm.aarch64.neon.vcmla.rot180.v2f32(<2 x float> %a, <2 x float> %b, <2 x float> %c)
ret <2 x float> %res
}
define <2 x float> @test_rot270_32x2(<2 x float> %a, <2 x float> %b, <2 x float> %c) {
entry:
; CHECK-LABEL: test_rot270_32x2
; CHECK: fcmla v{{[0-9]+}}.2s, v{{[0-9]+}}.2s, v{{[0-9]+}}.2s, #270
;
%res = tail call <2 x float> @llvm.aarch64.neon.vcmla.rot270.v2f32(<2 x float> %a, <2 x float> %b, <2 x float> %c)
ret <2 x float> %res
}
define <8 x half> @test_16x8(<8 x half> %a, <8 x half> %b, <8 x half> %c) {
entry:
; CHECK-LABEL: test_16x8
; CHECK: fcmla v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, #0
;
%res = tail call <8 x half> @llvm.aarch64.neon.vcmla.rot0.v8f16(<8 x half> %a, <8 x half> %b, <8 x half> %c)
ret <8 x half> %res
}
define <8 x half> @test_16x8_lane_0(<8 x half> %a, <8 x half> %b, <8 x half> %c) {
entry:
; CHECK-LABEL: test_16x8_lane_0
; CHECK: fcmla v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, v{{[0-9]+}}.h[0], #0
;
%c.cast = bitcast <8 x half> %c to <4 x i32>
%c.dup = shufflevector <4 x i32> %c.cast , <4 x i32> undef, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
%c.res = bitcast <4 x i32> %c.dup to <8 x half>
%res = tail call <8 x half> @llvm.aarch64.neon.vcmla.rot0.v8f16(<8 x half> %a, <8 x half> %b, <8 x half> %c.res)
ret <8 x half> %res
}
define <8 x half> @test_rot90_16x8(<8 x half> %a, <8 x half> %b, <8 x half> %c) {
entry:
; CHECK-LABEL: test_rot90_16x8
; CHECK: fcmla v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, #90
;
%res = tail call <8 x half> @llvm.aarch64.neon.vcmla.rot90.v8f16(<8 x half> %a, <8 x half> %b, <8 x half> %c)
ret <8 x half> %res
}
define <8 x half> @test_rot90_16x8_lane_1(<8 x half> %a, <8 x half> %b, <8 x half> %c) {
entry:
; CHECK-LABEL: test_rot90_16x8_lane_1
; CHECK: fcmla v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, v{{[0-9]+}}.h[1], #90
;
%c.cast = bitcast <8 x half> %c to <4 x i32>
%c.dup = shufflevector <4 x i32> %c.cast , <4 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
%c.res = bitcast <4 x i32> %c.dup to <8 x half>
%res = tail call <8 x half> @llvm.aarch64.neon.vcmla.rot90.v8f16(<8 x half> %a, <8 x half> %b, <8 x half> %c.res)
ret <8 x half> %res
}
define <8 x half> @test_rot180_16x8(<8 x half> %a, <8 x half> %b, <8 x half> %c) {
entry:
; CHECK-LABEL: test_rot180_16x8
; CHECK: fcmla v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, #180
;
%res = tail call <8 x half> @llvm.aarch64.neon.vcmla.rot180.v8f16(<8 x half> %a, <8 x half> %b, <8 x half> %c)
ret <8 x half> %res
}
define <8 x half> @test_rot180_16x8_lane_1(<8 x half> %a, <8 x half> %b, <8 x half> %c) {
entry:
; CHECK-LABEL: test_rot180_16x8_lane_1
; CHECK: fcmla v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, v{{[0-9]+}}.h[1], #180
;
%c.cast = bitcast <8 x half> %c to <4 x i32>
%c.dup = shufflevector <4 x i32> %c.cast , <4 x i32> undef, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
%c.res = bitcast <4 x i32> %c.dup to <8 x half>
%res = tail call <8 x half> @llvm.aarch64.neon.vcmla.rot180.v8f16(<8 x half> %a, <8 x half> %b, <8 x half> %c.res)
ret <8 x half> %res
}
define <8 x half> @test_rot270_16x8(<8 x half> %a, <8 x half> %b, <8 x half> %c) {
entry:
; CHECK-LABEL: test_rot270_16x8
; CHECK: fcmla v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, #270
;
%res = tail call <8 x half> @llvm.aarch64.neon.vcmla.rot270.v8f16(<8 x half> %a, <8 x half> %b, <8 x half> %c)
ret <8 x half> %res
}
define <8 x half> @test_rot270_16x8_lane_0(<8 x half> %a, <8 x half> %b, <8 x half> %c) {
entry:
; CHECK-LABEL: test_rot270_16x8_lane_0
; CHECK: fcmla v{{[0-9]+}}.8h, v{{[0-9]+}}.8h, v{{[0-9]+}}.h[0], #270
;
%c.cast = bitcast <8 x half> %c to <4 x i32>
%c.dup = shufflevector <4 x i32> %c.cast , <4 x i32> undef, <4 x i32> <i32 0, i32 0, i32 0, i32 0>
%c.res = bitcast <4 x i32> %c.dup to <8 x half>
%res = tail call <8 x half> @llvm.aarch64.neon.vcmla.rot270.v8f16(<8 x half> %a, <8 x half> %b, <8 x half> %c.res)
ret <8 x half> %res
}
define <4 x float> @test_32x4(<4 x float> %a, <4 x float> %b, <4 x float> %c) {
entry:
; CHECK-LABEL: test_32x4
; CHECK: fcmla v{{[0-9]+}}.4s, v{{[0-9]+}}.4s, v{{[0-9]+}}.4s, #0
;
%res = tail call <4 x float> @llvm.aarch64.neon.vcmla.rot0.v4f32(<4 x float> %a, <4 x float> %b, <4 x float> %c)
ret <4 x float> %res
}
define <4 x float> @test_32x4_lane_0(<4 x float> %a, <4 x float> %b, <4 x float> %c) {
entry:
; CHECK-LABEL: test_32x4_lane_0
; CHECK: fcmla v{{[0-9]+}}.4s, v{{[0-9]+}}.4s, v{{[0-9]+}}.s[0], #0
;
%c.cast = bitcast <4 x float> %c to <2 x i64>
%c.dup = shufflevector <2 x i64> %c.cast , <2 x i64> undef, <2 x i32> <i32 0, i32 0>
%c.res = bitcast <2 x i64> %c.dup to <4 x float>
%res = tail call <4 x float> @llvm.aarch64.neon.vcmla.rot0.v4f32(<4 x float> %a, <4 x float> %b, <4 x float> %c.res)
ret <4 x float> %res
}
define <4 x float> @test_rot90_32x4(<4 x float> %a, <4 x float> %b, <4 x float> %c) {
entry:
; CHECK-LABEL: test_rot90_32x4
; CHECK: fcmla v{{[0-9]+}}.4s, v{{[0-9]+}}.4s, v{{[0-9]+}}.4s, #90
;
%res = tail call <4 x float> @llvm.aarch64.neon.vcmla.rot90.v4f32(<4 x float> %a, <4 x float> %b, <4 x float> %c)
ret <4 x float> %res
}
define <4 x float> @test_rot180_32x4(<4 x float> %a, <4 x float> %b, <4 x float> %c) {
entry:
; CHECK-LABEL: test_rot180_32x4
; CHECK: fcmla v{{[0-9]+}}.4s, v{{[0-9]+}}.4s, v{{[0-9]+}}.4s, #180
;
%res = tail call <4 x float> @llvm.aarch64.neon.vcmla.rot180.v4f32(<4 x float> %a, <4 x float> %b, <4 x float> %c)
ret <4 x float> %res
}
define <4 x float> @test_rot270_32x4(<4 x float> %a, <4 x float> %b, <4 x float> %c) {
entry:
; CHECK-LABEL: test_rot270_32x4
; CHECK: fcmla v{{[0-9]+}}.4s, v{{[0-9]+}}.4s, v{{[0-9]+}}.4s, #270
;
%res = tail call <4 x float> @llvm.aarch64.neon.vcmla.rot270.v4f32(<4 x float> %a, <4 x float> %b, <4 x float> %c)
ret <4 x float> %res
}
define <2 x double> @test_64x2(<2 x double> %a, <2 x double> %b, <2 x double> %c) {
entry:
; CHECK-LABEL: test_64x2
; CHECK: fcmla v{{[0-9]+}}.2d, v{{[0-9]+}}.2d, v{{[0-9]+}}.2d, #0
;
%res = tail call <2 x double> @llvm.aarch64.neon.vcmla.rot0.v2f64(<2 x double> %a, <2 x double> %b, <2 x double> %c)
ret <2 x double> %res
}
define <2 x double> @test_rot90_64x2(<2 x double> %a, <2 x double> %b, <2 x double> %c) {
entry:
; CHECK-LABEL: test_rot90_64x2
; CHECK: fcmla v{{[0-9]+}}.2d, v{{[0-9]+}}.2d, v{{[0-9]+}}.2d, #90
;
%res = tail call <2 x double> @llvm.aarch64.neon.vcmla.rot90.v2f64(<2 x double> %a, <2 x double> %b, <2 x double> %c)
ret <2 x double> %res
}
define <2 x double> @test_rot180_64x2(<2 x double> %a, <2 x double> %b, <2 x double> %c) {
entry:
; CHECK-LABEL: test_rot180_64x2
; CHECK: fcmla v{{[0-9]+}}.2d, v{{[0-9]+}}.2d, v{{[0-9]+}}.2d, #180
;
%res = tail call <2 x double> @llvm.aarch64.neon.vcmla.rot180.v2f64(<2 x double> %a, <2 x double> %b, <2 x double> %c)
ret <2 x double> %res
}
define <2 x double> @test_rot270_64x2(<2 x double> %a, <2 x double> %b, <2 x double> %c) {
entry:
; CHECK-LABEL: test_rot270_64x2
; CHECK: fcmla v{{[0-9]+}}.2d, v{{[0-9]+}}.2d, v{{[0-9]+}}.2d, #270
;
%res = tail call <2 x double> @llvm.aarch64.neon.vcmla.rot270.v2f64(<2 x double> %a, <2 x double> %b, <2 x double> %c)
ret <2 x double> %res
}
declare <4 x half> @llvm.aarch64.neon.vcmla.rot0.v4f16(<4 x half>, <4 x half>, <4 x half>)
declare <4 x half> @llvm.aarch64.neon.vcmla.rot90.v4f16(<4 x half>, <4 x half>, <4 x half>)
declare <4 x half> @llvm.aarch64.neon.vcmla.rot180.v4f16(<4 x half>, <4 x half>, <4 x half>)
declare <4 x half> @llvm.aarch64.neon.vcmla.rot270.v4f16(<4 x half>, <4 x half>, <4 x half>)
declare <8 x half> @llvm.aarch64.neon.vcmla.rot0.v8f16(<8 x half>, <8 x half>, <8 x half>)
declare <8 x half> @llvm.aarch64.neon.vcmla.rot90.v8f16(<8 x half>, <8 x half>, <8 x half>)
declare <8 x half> @llvm.aarch64.neon.vcmla.rot180.v8f16(<8 x half>, <8 x half>, <8 x half>)
declare <8 x half> @llvm.aarch64.neon.vcmla.rot270.v8f16(<8 x half>, <8 x half>, <8 x half>)
declare <2 x float> @llvm.aarch64.neon.vcmla.rot0.v2f32(<2 x float>, <2 x float>, <2 x float>)
declare <2 x float> @llvm.aarch64.neon.vcmla.rot90.v2f32(<2 x float>, <2 x float>, <2 x float>)
declare <2 x float> @llvm.aarch64.neon.vcmla.rot180.v2f32(<2 x float>, <2 x float>, <2 x float>)
declare <2 x float> @llvm.aarch64.neon.vcmla.rot270.v2f32(<2 x float>, <2 x float>, <2 x float>)
declare <4 x float> @llvm.aarch64.neon.vcmla.rot0.v4f32(<4 x float>, <4 x float>, <4 x float>)
declare <4 x float> @llvm.aarch64.neon.vcmla.rot90.v4f32(<4 x float>, <4 x float>, <4 x float>)
declare <4 x float> @llvm.aarch64.neon.vcmla.rot180.v4f32(<4 x float>, <4 x float>, <4 x float>)
declare <4 x float> @llvm.aarch64.neon.vcmla.rot270.v4f32(<4 x float>, <4 x float>, <4 x float>)
declare <2 x double> @llvm.aarch64.neon.vcmla.rot0.v2f64(<2 x double>, <2 x double>, <2 x double>)
declare <2 x double> @llvm.aarch64.neon.vcmla.rot90.v2f64(<2 x double>, <2 x double>, <2 x double>)
declare <2 x double> @llvm.aarch64.neon.vcmla.rot180.v2f64(<2 x double>, <2 x double>, <2 x double>)
declare <2 x double> @llvm.aarch64.neon.vcmla.rot270.v2f64(<2 x double>, <2 x double>, <2 x double>)
|
