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
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
|
// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +neon -target-feature +bf16 -target-feature +fp8 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,sroa | FileCheck %s
// RUN: %clang_cc1 -x c++ -triple aarch64-none-linux-gnu -target-feature +neon -target-feature +bf16 -target-feature +fp8 -disable-O0-optnone -Werror -Wall -emit-llvm -o - %s | opt -S -p mem2reg,sroa | FileCheck %s -check-prefix CHECK-CXX
// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +neon -target-feature +bf16 -target-feature +fp8 -S -O3 -o /dev/null %s
// REQUIRES: aarch64-registered-target
#include <arm_neon.h>
// CHECK-LABEL: define dso_local <8 x bfloat> @test_vcvt1_bf16_mf8_fpm(
// CHECK-SAME: <8 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT1_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl1.v8bf16.v8i8(<8 x i8> [[OP]])
// CHECK-NEXT: ret <8 x bfloat> [[VBFCVT1_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x bfloat> @_Z23test_vcvt1_bf16_mf8_fpm13__Mfloat8x8_tm(
// CHECK-CXX-SAME: <8 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT1_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl1.v8bf16.v8i8(<8 x i8> [[OP]])
// CHECK-CXX-NEXT: ret <8 x bfloat> [[VBFCVT1_I]]
//
bfloat16x8_t test_vcvt1_bf16_mf8_fpm(mfloat8x8_t op, fpm_t fpm) {
return vcvt1_bf16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x bfloat> @test_vcvt1_low_bf16_mf8_fpm(
// CHECK-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[TMP0:%.*]] = call <8 x i8> @llvm.vector.extract.v8i8.v16i8(<16 x i8> [[OP]], i64 0)
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT1_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl1.v8bf16.v8i8(<8 x i8> [[TMP0]])
// CHECK-NEXT: ret <8 x bfloat> [[VBFCVT1_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x bfloat> @_Z27test_vcvt1_low_bf16_mf8_fpm14__Mfloat8x16_tm(
// CHECK-CXX-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: [[TMP0:%.*]] = call <8 x i8> @llvm.vector.extract.v8i8.v16i8(<16 x i8> [[OP]], i64 0)
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT1_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl1.v8bf16.v8i8(<8 x i8> [[TMP0]])
// CHECK-CXX-NEXT: ret <8 x bfloat> [[VBFCVT1_I]]
//
bfloat16x8_t test_vcvt1_low_bf16_mf8_fpm(mfloat8x16_t op, fpm_t fpm) {
return vcvt1_low_bf16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x bfloat> @test_vcvt2_bf16_mf8_fpm(
// CHECK-SAME: <8 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT2_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl2.v8bf16.v8i8(<8 x i8> [[OP]])
// CHECK-NEXT: ret <8 x bfloat> [[VBFCVT2_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x bfloat> @_Z23test_vcvt2_bf16_mf8_fpm13__Mfloat8x8_tm(
// CHECK-CXX-SAME: <8 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT2_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl2.v8bf16.v8i8(<8 x i8> [[OP]])
// CHECK-CXX-NEXT: ret <8 x bfloat> [[VBFCVT2_I]]
//
bfloat16x8_t test_vcvt2_bf16_mf8_fpm(mfloat8x8_t op, fpm_t fpm) {
return vcvt2_bf16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x bfloat> @test_vcvt2_low_bf16_mf8_fpm(
// CHECK-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[TMP0:%.*]] = call <8 x i8> @llvm.vector.extract.v8i8.v16i8(<16 x i8> [[OP]], i64 0)
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT2_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl2.v8bf16.v8i8(<8 x i8> [[TMP0]])
// CHECK-NEXT: ret <8 x bfloat> [[VBFCVT2_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x bfloat> @_Z27test_vcvt2_low_bf16_mf8_fpm14__Mfloat8x16_tm(
// CHECK-CXX-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: [[TMP0:%.*]] = call <8 x i8> @llvm.vector.extract.v8i8.v16i8(<16 x i8> [[OP]], i64 0)
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT2_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl2.v8bf16.v8i8(<8 x i8> [[TMP0]])
// CHECK-CXX-NEXT: ret <8 x bfloat> [[VBFCVT2_I]]
//
bfloat16x8_t test_vcvt2_low_bf16_mf8_fpm(mfloat8x16_t op, fpm_t fpm) {
return vcvt2_low_bf16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x bfloat> @test_vcvt1_high_bf16_mf8_fpm(
// CHECK-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT1_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl1.v8bf16.v16i8(<16 x i8> [[OP]])
// CHECK-NEXT: ret <8 x bfloat> [[VBFCVT1_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x bfloat> @_Z28test_vcvt1_high_bf16_mf8_fpm14__Mfloat8x16_tm(
// CHECK-CXX-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT1_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl1.v8bf16.v16i8(<16 x i8> [[OP]])
// CHECK-CXX-NEXT: ret <8 x bfloat> [[VBFCVT1_I]]
//
bfloat16x8_t test_vcvt1_high_bf16_mf8_fpm(mfloat8x16_t op, fpm_t fpm) {
return vcvt1_high_bf16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x bfloat> @test_vcvt2_high_bf16_mf8_fpm(
// CHECK-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT2_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl2.v8bf16.v16i8(<16 x i8> [[OP]])
// CHECK-NEXT: ret <8 x bfloat> [[VBFCVT2_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x bfloat> @_Z28test_vcvt2_high_bf16_mf8_fpm14__Mfloat8x16_tm(
// CHECK-CXX-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT2_I:%.*]] = call <8 x bfloat> @llvm.aarch64.neon.fp8.cvtl2.v8bf16.v16i8(<16 x i8> [[OP]])
// CHECK-CXX-NEXT: ret <8 x bfloat> [[VBFCVT2_I]]
//
bfloat16x8_t test_vcvt2_high_bf16_mf8_fpm(mfloat8x16_t op, fpm_t fpm) {
return vcvt2_high_bf16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x half> @test_vcvt1_f16_mf8_fpm(
// CHECK-SAME: <8 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT1_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl1.v8f16.v8i8(<8 x i8> [[OP]])
// CHECK-NEXT: ret <8 x half> [[VBFCVT1_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x half> @_Z22test_vcvt1_f16_mf8_fpm13__Mfloat8x8_tm(
// CHECK-CXX-SAME: <8 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT1_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl1.v8f16.v8i8(<8 x i8> [[OP]])
// CHECK-CXX-NEXT: ret <8 x half> [[VBFCVT1_I]]
//
float16x8_t test_vcvt1_f16_mf8_fpm(mfloat8x8_t op, fpm_t fpm) {
return vcvt1_f16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x half> @test_vcvt1_low_f16_mf8_fpm(
// CHECK-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[TMP0:%.*]] = call <8 x i8> @llvm.vector.extract.v8i8.v16i8(<16 x i8> [[OP]], i64 0)
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT1_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl1.v8f16.v8i8(<8 x i8> [[TMP0]])
// CHECK-NEXT: ret <8 x half> [[VBFCVT1_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x half> @_Z26test_vcvt1_low_f16_mf8_fpm14__Mfloat8x16_tm(
// CHECK-CXX-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: [[TMP0:%.*]] = call <8 x i8> @llvm.vector.extract.v8i8.v16i8(<16 x i8> [[OP]], i64 0)
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT1_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl1.v8f16.v8i8(<8 x i8> [[TMP0]])
// CHECK-CXX-NEXT: ret <8 x half> [[VBFCVT1_I]]
//
float16x8_t test_vcvt1_low_f16_mf8_fpm(mfloat8x16_t op, fpm_t fpm) {
return vcvt1_low_f16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x half> @test_vcvt2_f16_mf8_fpm(
// CHECK-SAME: <8 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT2_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl2.v8f16.v8i8(<8 x i8> [[OP]])
// CHECK-NEXT: ret <8 x half> [[VBFCVT2_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x half> @_Z22test_vcvt2_f16_mf8_fpm13__Mfloat8x8_tm(
// CHECK-CXX-SAME: <8 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT2_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl2.v8f16.v8i8(<8 x i8> [[OP]])
// CHECK-CXX-NEXT: ret <8 x half> [[VBFCVT2_I]]
//
float16x8_t test_vcvt2_f16_mf8_fpm(mfloat8x8_t op, fpm_t fpm) {
return vcvt2_f16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x half> @test_vcvt2_low_f16_mf8_fpm(
// CHECK-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[TMP0:%.*]] = call <8 x i8> @llvm.vector.extract.v8i8.v16i8(<16 x i8> [[OP]], i64 0)
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT2_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl2.v8f16.v8i8(<8 x i8> [[TMP0]])
// CHECK-NEXT: ret <8 x half> [[VBFCVT2_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x half> @_Z26test_vcvt2_low_f16_mf8_fpm14__Mfloat8x16_tm(
// CHECK-CXX-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: [[TMP0:%.*]] = call <8 x i8> @llvm.vector.extract.v8i8.v16i8(<16 x i8> [[OP]], i64 0)
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT2_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl2.v8f16.v8i8(<8 x i8> [[TMP0]])
// CHECK-CXX-NEXT: ret <8 x half> [[VBFCVT2_I]]
//
float16x8_t test_vcvt2_low_f16_mf8_fpm(mfloat8x16_t op, fpm_t fpm) {
return vcvt2_low_f16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x half> @test_vcvt1_high_f16_mf8_fpm(
// CHECK-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT1_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl1.v8f16.v16i8(<16 x i8> [[OP]])
// CHECK-NEXT: ret <8 x half> [[VBFCVT1_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x half> @_Z27test_vcvt1_high_f16_mf8_fpm14__Mfloat8x16_tm(
// CHECK-CXX-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT1_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl1.v8f16.v16i8(<16 x i8> [[OP]])
// CHECK-CXX-NEXT: ret <8 x half> [[VBFCVT1_I]]
//
float16x8_t test_vcvt1_high_f16_mf8_fpm(mfloat8x16_t op, fpm_t fpm) {
return vcvt1_high_f16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x half> @test_vcvt2_high_f16_mf8_fpm(
// CHECK-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VBFCVT2_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl2.v8f16.v16i8(<16 x i8> [[OP]])
// CHECK-NEXT: ret <8 x half> [[VBFCVT2_I]]
//
// CHECK-CXX-LABEL: define dso_local noundef <8 x half> @_Z27test_vcvt2_high_f16_mf8_fpm14__Mfloat8x16_tm(
// CHECK-CXX-SAME: <16 x i8> [[OP:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VBFCVT2_I:%.*]] = call <8 x half> @llvm.aarch64.neon.fp8.cvtl2.v8f16.v16i8(<16 x i8> [[OP]])
// CHECK-CXX-NEXT: ret <8 x half> [[VBFCVT2_I]]
//
float16x8_t test_vcvt2_high_f16_mf8_fpm(mfloat8x16_t op, fpm_t fpm) {
return vcvt2_high_f16_mf8_fpm(op, fpm);
}
// CHECK-LABEL: define dso_local <8 x i8> @test_vcvt_mf8_f32_fpm(
// CHECK-SAME: <4 x float> noundef [[VN:%.*]], <4 x float> noundef [[VM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VFCVTN_I:%.*]] = call <8 x i8> @llvm.aarch64.neon.fp8.fcvtn.v8i8.v4f32(<4 x float> [[VN]], <4 x float> [[VM]])
// CHECK-NEXT: ret <8 x i8> [[VFCVTN_I]]
//
// CHECK-CXX-LABEL: define dso_local <8 x i8> @_Z21test_vcvt_mf8_f32_fpm13__Float32x4_tS_m(
// CHECK-CXX-SAME: <4 x float> noundef [[VN:%.*]], <4 x float> noundef [[VM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VFCVTN_I:%.*]] = call <8 x i8> @llvm.aarch64.neon.fp8.fcvtn.v8i8.v4f32(<4 x float> [[VN]], <4 x float> [[VM]])
// CHECK-CXX-NEXT: ret <8 x i8> [[VFCVTN_I]]
//
mfloat8x8_t test_vcvt_mf8_f32_fpm(float32x4_t vn, float32x4_t vm, fpm_t fpm) {
return vcvt_mf8_f32_fpm(vn, vm, fpm);
}
// CHECK-LABEL: define dso_local <16 x i8> @test_vcvt_high_mf8_f32_fpm(
// CHECK-SAME: <8 x i8> [[VD:%.*]], <4 x float> noundef [[VN:%.*]], <4 x float> noundef [[VM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[TMP0:%.*]] = call <16 x i8> @llvm.vector.insert.v16i8.v8i8(<16 x i8> poison, <8 x i8> [[VD]], i64 0)
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VFCVTN2_I:%.*]] = call <16 x i8> @llvm.aarch64.neon.fp8.fcvtn2.v16i8.v4f32(<16 x i8> [[TMP0]], <4 x float> [[VN]], <4 x float> [[VM]])
// CHECK-NEXT: ret <16 x i8> [[VFCVTN2_I]]
//
// CHECK-CXX-LABEL: define dso_local <16 x i8> @_Z26test_vcvt_high_mf8_f32_fpm13__Mfloat8x8_t13__Float32x4_tS0_m(
// CHECK-CXX-SAME: <8 x i8> [[VD:%.*]], <4 x float> noundef [[VN:%.*]], <4 x float> noundef [[VM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: [[TMP0:%.*]] = call <16 x i8> @llvm.vector.insert.v16i8.v8i8(<16 x i8> poison, <8 x i8> [[VD]], i64 0)
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VFCVTN2_I:%.*]] = call <16 x i8> @llvm.aarch64.neon.fp8.fcvtn2.v16i8.v4f32(<16 x i8> [[TMP0]], <4 x float> [[VN]], <4 x float> [[VM]])
// CHECK-CXX-NEXT: ret <16 x i8> [[VFCVTN2_I]]
//
mfloat8x16_t test_vcvt_high_mf8_f32_fpm(mfloat8x8_t vd, float32x4_t vn,
float32x4_t vm, fpm_t fpm) {
return vcvt_high_mf8_f32_fpm(vd, vn, vm, fpm);
}
// CHECK-LABEL: define dso_local <8 x i8> @test_vcvt_mf8_f16_fpm(
// CHECK-SAME: <4 x half> noundef [[VN:%.*]], <4 x half> noundef [[VM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[TMP0:%.*]] = bitcast <4 x half> [[VN]] to <4 x i16>
// CHECK-NEXT: [[TMP1:%.*]] = bitcast <4 x half> [[VM]] to <4 x i16>
// CHECK-NEXT: [[TMP2:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-NEXT: [[TMP3:%.*]] = bitcast <4 x i16> [[TMP1]] to <8 x i8>
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VFCVTN_I:%.*]] = bitcast <8 x i8> [[TMP2]] to <4 x half>
// CHECK-NEXT: [[VFCVTN1_I:%.*]] = bitcast <8 x i8> [[TMP3]] to <4 x half>
// CHECK-NEXT: [[VFCVTN2_I:%.*]] = call <8 x i8> @llvm.aarch64.neon.fp8.fcvtn.v8i8.v4f16(<4 x half> [[VFCVTN_I]], <4 x half> [[VFCVTN1_I]])
// CHECK-NEXT: ret <8 x i8> [[VFCVTN2_I]]
//
// CHECK-CXX-LABEL: define dso_local <8 x i8> @_Z21test_vcvt_mf8_f16_fpm13__Float16x4_tS_m(
// CHECK-CXX-SAME: <4 x half> noundef [[VN:%.*]], <4 x half> noundef [[VM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: [[TMP0:%.*]] = bitcast <4 x half> [[VN]] to <4 x i16>
// CHECK-CXX-NEXT: [[TMP1:%.*]] = bitcast <4 x half> [[VM]] to <4 x i16>
// CHECK-CXX-NEXT: [[TMP2:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-CXX-NEXT: [[TMP3:%.*]] = bitcast <4 x i16> [[TMP1]] to <8 x i8>
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VFCVTN_I:%.*]] = bitcast <8 x i8> [[TMP2]] to <4 x half>
// CHECK-CXX-NEXT: [[VFCVTN1_I:%.*]] = bitcast <8 x i8> [[TMP3]] to <4 x half>
// CHECK-CXX-NEXT: [[VFCVTN2_I:%.*]] = call <8 x i8> @llvm.aarch64.neon.fp8.fcvtn.v8i8.v4f16(<4 x half> [[VFCVTN_I]], <4 x half> [[VFCVTN1_I]])
// CHECK-CXX-NEXT: ret <8 x i8> [[VFCVTN2_I]]
//
mfloat8x8_t test_vcvt_mf8_f16_fpm(float16x4_t vn, float16x4_t vm, fpm_t fpm) {
return vcvt_mf8_f16_fpm(vn, vm, fpm);
}
// CHECK-LABEL: define dso_local <16 x i8> @test_vcvtq_mf8_f16_fpm(
// CHECK-SAME: <8 x half> noundef [[VN:%.*]], <8 x half> noundef [[VM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[TMP0:%.*]] = bitcast <8 x half> [[VN]] to <8 x i16>
// CHECK-NEXT: [[TMP1:%.*]] = bitcast <8 x half> [[VM]] to <8 x i16>
// CHECK-NEXT: [[TMP2:%.*]] = bitcast <8 x i16> [[TMP0]] to <16 x i8>
// CHECK-NEXT: [[TMP3:%.*]] = bitcast <8 x i16> [[TMP1]] to <16 x i8>
// CHECK-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-NEXT: [[VFCVTN_I:%.*]] = bitcast <16 x i8> [[TMP2]] to <8 x half>
// CHECK-NEXT: [[VFCVTN1_I:%.*]] = bitcast <16 x i8> [[TMP3]] to <8 x half>
// CHECK-NEXT: [[VFCVTN2_I:%.*]] = call <16 x i8> @llvm.aarch64.neon.fp8.fcvtn.v16i8.v8f16(<8 x half> [[VFCVTN_I]], <8 x half> [[VFCVTN1_I]])
// CHECK-NEXT: ret <16 x i8> [[VFCVTN2_I]]
//
// CHECK-CXX-LABEL: define dso_local <16 x i8> @_Z22test_vcvtq_mf8_f16_fpm13__Float16x8_tS_m(
// CHECK-CXX-SAME: <8 x half> noundef [[VN:%.*]], <8 x half> noundef [[VM:%.*]], i64 noundef [[FPM:%.*]]) #[[ATTR0]] {
// CHECK-CXX-NEXT: [[ENTRY:.*:]]
// CHECK-CXX-NEXT: [[TMP0:%.*]] = bitcast <8 x half> [[VN]] to <8 x i16>
// CHECK-CXX-NEXT: [[TMP1:%.*]] = bitcast <8 x half> [[VM]] to <8 x i16>
// CHECK-CXX-NEXT: [[TMP2:%.*]] = bitcast <8 x i16> [[TMP0]] to <16 x i8>
// CHECK-CXX-NEXT: [[TMP3:%.*]] = bitcast <8 x i16> [[TMP1]] to <16 x i8>
// CHECK-CXX-NEXT: call void @llvm.aarch64.set.fpmr(i64 [[FPM]])
// CHECK-CXX-NEXT: [[VFCVTN_I:%.*]] = bitcast <16 x i8> [[TMP2]] to <8 x half>
// CHECK-CXX-NEXT: [[VFCVTN1_I:%.*]] = bitcast <16 x i8> [[TMP3]] to <8 x half>
// CHECK-CXX-NEXT: [[VFCVTN2_I:%.*]] = call <16 x i8> @llvm.aarch64.neon.fp8.fcvtn.v16i8.v8f16(<8 x half> [[VFCVTN_I]], <8 x half> [[VFCVTN1_I]])
// CHECK-CXX-NEXT: ret <16 x i8> [[VFCVTN2_I]]
//
mfloat8x16_t test_vcvtq_mf8_f16_fpm(float16x8_t vn, float16x8_t vm, fpm_t fpm) {
return vcvtq_mf8_f16_fpm(vn, vm, fpm);
}
|
