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
|
// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
// RUN: %clang_cc1 \
// RUN: -triple aarch64 -target-feature +neon -target-feature +bf16 \
// RUN: -disable-O0-optnone -emit-llvm -o - %s \
// RUN: | opt -S -passes=mem2reg,sroa \
// RUN: | FileCheck --check-prefixes=CHECK,CHECK-A64 %s
// RUN: %clang_cc1 \
// RUN: -triple armv8.6a-arm-none-eabi -target-feature +neon \
// RUN: -target-feature +bf16 -mfloat-abi hard \
// RUN: -disable-O0-optnone -emit-llvm -o - %s \
// RUN: | opt -S -passes=mem2reg,sroa \
// RUN: | FileCheck --check-prefixes=CHECK,CHECK-A32-HARDFP %s
// RUN: %clang_cc1 \
// RUN: -triple armv8.6a-arm-none-eabi -target-feature +neon \
// RUN: -target-feature +bf16 -mfloat-abi softfp \
// RUN: -disable-O0-optnone -emit-llvm -o - %s \
// RUN: | opt -S -passes=mem2reg,sroa \
// RUN: | FileCheck --check-prefixes=CHECK,CHECK-A32-SOFTFP %s
// REQUIRES: arm-registered-target
// REQUIRES: aarch64-registered-target
#include <arm_neon.h>
// CHECK-A64-LABEL: @test_vcvt_f32_bf16(
// CHECK-A64-NEXT: entry:
// CHECK-A64-NEXT: [[TMP0:%.*]] = bitcast <4 x bfloat> [[A:%.*]] to <4 x i16>
// CHECK-A64-NEXT: [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-A64-NEXT: [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
// CHECK-A64-NEXT: [[TMP3:%.*]] = zext <4 x i16> [[TMP2]] to <4 x i32>
// CHECK-A64-NEXT: [[VSHLL_N_I:%.*]] = shl <4 x i32> [[TMP3]], splat (i32 16)
// CHECK-A64-NEXT: [[TMP4:%.*]] = bitcast <4 x i32> [[VSHLL_N_I]] to <4 x float>
// CHECK-A64-NEXT: ret <4 x float> [[TMP4]]
//
// CHECK-A32-HARDFP-LABEL: @test_vcvt_f32_bf16(
// CHECK-A32-HARDFP-NEXT: entry:
// CHECK-A32-HARDFP-NEXT: [[TMP0:%.*]] = bitcast <4 x bfloat> [[A:%.*]] to <4 x i16>
// CHECK-A32-HARDFP-NEXT: [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-A32-HARDFP-NEXT: [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
// CHECK-A32-HARDFP-NEXT: [[TMP3:%.*]] = zext <4 x i16> [[TMP2]] to <4 x i32>
// CHECK-A32-HARDFP-NEXT: [[VSHLL_N_I:%.*]] = shl <4 x i32> [[TMP3]], splat (i32 16)
// CHECK-A32-HARDFP-NEXT: [[TMP4:%.*]] = bitcast <4 x i32> [[VSHLL_N_I]] to <4 x float>
// CHECK-A32-HARDFP-NEXT: ret <4 x float> [[TMP4]]
//
// CHECK-A32-SOFTFP-LABEL: @test_vcvt_f32_bf16(
// CHECK-A32-SOFTFP-NEXT: entry:
// CHECK-A32-SOFTFP-NEXT: [[TMP0:%.*]] = bitcast <2 x i32> [[A_COERCE:%.*]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP1:%.*]] = bitcast <4 x bfloat> [[TMP0]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP2:%.*]] = bitcast <2 x i32> [[TMP1]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP3:%.*]] = bitcast <4 x bfloat> [[TMP2]] to <4 x i16>
// CHECK-A32-SOFTFP-NEXT: [[TMP4:%.*]] = bitcast <4 x i16> [[TMP3]] to <8 x i8>
// CHECK-A32-SOFTFP-NEXT: [[TMP5:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
// CHECK-A32-SOFTFP-NEXT: [[TMP6:%.*]] = zext <4 x i16> [[TMP5]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[VSHLL_N_I:%.*]] = shl <4 x i32> [[TMP6]], splat (i32 16)
// CHECK-A32-SOFTFP-NEXT: [[TMP7:%.*]] = bitcast <4 x i32> [[VSHLL_N_I]] to <4 x float>
// CHECK-A32-SOFTFP-NEXT: ret <4 x float> [[TMP7]]
//
float32x4_t test_vcvt_f32_bf16(bfloat16x4_t a) {
return vcvt_f32_bf16(a);
}
// CHECK-A64-LABEL: @test_vcvtq_low_f32_bf16(
// CHECK-A64-NEXT: entry:
// CHECK-A64-NEXT: [[SHUFFLE_I:%.*]] = shufflevector <8 x bfloat> [[A:%.*]], <8 x bfloat> [[A]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
// CHECK-A64-NEXT: [[TMP0:%.*]] = bitcast <4 x bfloat> [[SHUFFLE_I]] to <4 x i16>
// CHECK-A64-NEXT: [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-A64-NEXT: [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
// CHECK-A64-NEXT: [[TMP3:%.*]] = zext <4 x i16> [[TMP2]] to <4 x i32>
// CHECK-A64-NEXT: [[VSHLL_N_I_I:%.*]] = shl <4 x i32> [[TMP3]], splat (i32 16)
// CHECK-A64-NEXT: [[TMP4:%.*]] = bitcast <4 x i32> [[VSHLL_N_I_I]] to <4 x float>
// CHECK-A64-NEXT: ret <4 x float> [[TMP4]]
//
// CHECK-A32-HARDFP-LABEL: @test_vcvtq_low_f32_bf16(
// CHECK-A32-HARDFP-NEXT: entry:
// CHECK-A32-HARDFP-NEXT: [[SHUFFLE_I:%.*]] = shufflevector <8 x bfloat> [[A:%.*]], <8 x bfloat> [[A]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
// CHECK-A32-HARDFP-NEXT: [[TMP0:%.*]] = bitcast <4 x bfloat> [[SHUFFLE_I]] to <4 x i16>
// CHECK-A32-HARDFP-NEXT: [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-A32-HARDFP-NEXT: [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
// CHECK-A32-HARDFP-NEXT: [[TMP3:%.*]] = zext <4 x i16> [[TMP2]] to <4 x i32>
// CHECK-A32-HARDFP-NEXT: [[VSHLL_N_I_I:%.*]] = shl <4 x i32> [[TMP3]], splat (i32 16)
// CHECK-A32-HARDFP-NEXT: [[TMP4:%.*]] = bitcast <4 x i32> [[VSHLL_N_I_I]] to <4 x float>
// CHECK-A32-HARDFP-NEXT: ret <4 x float> [[TMP4]]
//
// CHECK-A32-SOFTFP-LABEL: @test_vcvtq_low_f32_bf16(
// CHECK-A32-SOFTFP-NEXT: entry:
// CHECK-A32-SOFTFP-NEXT: [[TMP0:%.*]] = bitcast <4 x i32> [[A_COERCE:%.*]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP1:%.*]] = bitcast <8 x bfloat> [[TMP0]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP2:%.*]] = bitcast <4 x i32> [[TMP1]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP3:%.*]] = bitcast <8 x bfloat> [[TMP2]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP4:%.*]] = bitcast <4 x i32> [[TMP3]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[SHUFFLE_I:%.*]] = shufflevector <8 x bfloat> [[TMP4]], <8 x bfloat> [[TMP4]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
// CHECK-A32-SOFTFP-NEXT: [[TMP5:%.*]] = bitcast <4 x bfloat> [[SHUFFLE_I]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP6:%.*]] = bitcast <2 x i32> [[TMP5]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP7:%.*]] = bitcast <4 x bfloat> [[TMP6]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP8:%.*]] = bitcast <2 x i32> [[TMP7]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP9:%.*]] = bitcast <4 x bfloat> [[TMP8]] to <4 x i16>
// CHECK-A32-SOFTFP-NEXT: [[TMP10:%.*]] = bitcast <4 x i16> [[TMP9]] to <8 x i8>
// CHECK-A32-SOFTFP-NEXT: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x i16>
// CHECK-A32-SOFTFP-NEXT: [[TMP12:%.*]] = zext <4 x i16> [[TMP11]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[VSHLL_N_I_I:%.*]] = shl <4 x i32> [[TMP12]], splat (i32 16)
// CHECK-A32-SOFTFP-NEXT: [[TMP13:%.*]] = bitcast <4 x i32> [[VSHLL_N_I_I]] to <4 x float>
// CHECK-A32-SOFTFP-NEXT: ret <4 x float> [[TMP13]]
//
float32x4_t test_vcvtq_low_f32_bf16(bfloat16x8_t a) {
return vcvtq_low_f32_bf16(a);
}
// CHECK-A64-LABEL: @test_vcvtq_high_f32_bf16(
// CHECK-A64-NEXT: entry:
// CHECK-A64-NEXT: [[SHUFFLE_I:%.*]] = shufflevector <8 x bfloat> [[A:%.*]], <8 x bfloat> [[A]], <4 x i32> <i32 4, i32 5, i32 6, i32 7>
// CHECK-A64-NEXT: [[TMP0:%.*]] = bitcast <4 x bfloat> [[SHUFFLE_I]] to <4 x i16>
// CHECK-A64-NEXT: [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-A64-NEXT: [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
// CHECK-A64-NEXT: [[TMP3:%.*]] = zext <4 x i16> [[TMP2]] to <4 x i32>
// CHECK-A64-NEXT: [[VSHLL_N_I_I:%.*]] = shl <4 x i32> [[TMP3]], splat (i32 16)
// CHECK-A64-NEXT: [[TMP4:%.*]] = bitcast <4 x i32> [[VSHLL_N_I_I]] to <4 x float>
// CHECK-A64-NEXT: ret <4 x float> [[TMP4]]
//
// CHECK-A32-HARDFP-LABEL: @test_vcvtq_high_f32_bf16(
// CHECK-A32-HARDFP-NEXT: entry:
// CHECK-A32-HARDFP-NEXT: [[SHUFFLE_I:%.*]] = shufflevector <8 x bfloat> [[A:%.*]], <8 x bfloat> [[A]], <4 x i32> <i32 4, i32 5, i32 6, i32 7>
// CHECK-A32-HARDFP-NEXT: [[TMP0:%.*]] = bitcast <4 x bfloat> [[SHUFFLE_I]] to <4 x i16>
// CHECK-A32-HARDFP-NEXT: [[TMP1:%.*]] = bitcast <4 x i16> [[TMP0]] to <8 x i8>
// CHECK-A32-HARDFP-NEXT: [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
// CHECK-A32-HARDFP-NEXT: [[TMP3:%.*]] = zext <4 x i16> [[TMP2]] to <4 x i32>
// CHECK-A32-HARDFP-NEXT: [[VSHLL_N_I_I:%.*]] = shl <4 x i32> [[TMP3]], splat (i32 16)
// CHECK-A32-HARDFP-NEXT: [[TMP4:%.*]] = bitcast <4 x i32> [[VSHLL_N_I_I]] to <4 x float>
// CHECK-A32-HARDFP-NEXT: ret <4 x float> [[TMP4]]
//
// CHECK-A32-SOFTFP-LABEL: @test_vcvtq_high_f32_bf16(
// CHECK-A32-SOFTFP-NEXT: entry:
// CHECK-A32-SOFTFP-NEXT: [[TMP0:%.*]] = bitcast <4 x i32> [[A_COERCE:%.*]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP1:%.*]] = bitcast <8 x bfloat> [[TMP0]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP2:%.*]] = bitcast <4 x i32> [[TMP1]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP3:%.*]] = bitcast <8 x bfloat> [[TMP2]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP4:%.*]] = bitcast <4 x i32> [[TMP3]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[SHUFFLE_I:%.*]] = shufflevector <8 x bfloat> [[TMP4]], <8 x bfloat> [[TMP4]], <4 x i32> <i32 4, i32 5, i32 6, i32 7>
// CHECK-A32-SOFTFP-NEXT: [[TMP5:%.*]] = bitcast <4 x bfloat> [[SHUFFLE_I]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP6:%.*]] = bitcast <2 x i32> [[TMP5]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP7:%.*]] = bitcast <4 x bfloat> [[TMP6]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP8:%.*]] = bitcast <2 x i32> [[TMP7]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP9:%.*]] = bitcast <4 x bfloat> [[TMP8]] to <4 x i16>
// CHECK-A32-SOFTFP-NEXT: [[TMP10:%.*]] = bitcast <4 x i16> [[TMP9]] to <8 x i8>
// CHECK-A32-SOFTFP-NEXT: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x i16>
// CHECK-A32-SOFTFP-NEXT: [[TMP12:%.*]] = zext <4 x i16> [[TMP11]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[VSHLL_N_I_I:%.*]] = shl <4 x i32> [[TMP12]], splat (i32 16)
// CHECK-A32-SOFTFP-NEXT: [[TMP13:%.*]] = bitcast <4 x i32> [[VSHLL_N_I_I]] to <4 x float>
// CHECK-A32-SOFTFP-NEXT: ret <4 x float> [[TMP13]]
//
float32x4_t test_vcvtq_high_f32_bf16(bfloat16x8_t a) {
return vcvtq_high_f32_bf16(a);
}
// CHECK-A64-LABEL: @test_vcvt_bf16_f32(
// CHECK-A64-NEXT: entry:
// CHECK-A64-NEXT: [[TMP0:%.*]] = bitcast <4 x float> [[A:%.*]] to <4 x i32>
// CHECK-A64-NEXT: [[TMP1:%.*]] = bitcast <4 x i32> [[TMP0]] to <16 x i8>
// CHECK-A64-NEXT: [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <4 x float>
// CHECK-A64-NEXT: [[TMP3:%.*]] = fptrunc <4 x float> [[TMP2]] to <4 x bfloat>
// CHECK-A64-NEXT: ret <4 x bfloat> [[TMP3]]
//
// CHECK-A32-HARDFP-LABEL: @test_vcvt_bf16_f32(
// CHECK-A32-HARDFP-NEXT: entry:
// CHECK-A32-HARDFP-NEXT: [[TMP0:%.*]] = bitcast <4 x float> [[A:%.*]] to <4 x i32>
// CHECK-A32-HARDFP-NEXT: [[TMP1:%.*]] = bitcast <4 x i32> [[TMP0]] to <16 x i8>
// CHECK-A32-HARDFP-NEXT: [[VCVTFP2BF_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <4 x float>
// CHECK-A32-HARDFP-NEXT: [[VCVTFP2BF1_I:%.*]] = call <4 x bfloat> @llvm.arm.neon.vcvtfp2bf.v4bf16(<4 x float> [[VCVTFP2BF_I]])
// CHECK-A32-HARDFP-NEXT: ret <4 x bfloat> [[VCVTFP2BF1_I]]
//
// CHECK-A32-SOFTFP-LABEL: @test_vcvt_bf16_f32(
// CHECK-A32-SOFTFP-NEXT: entry:
// CHECK-A32-SOFTFP-NEXT: [[TMP0:%.*]] = bitcast <4 x float> [[A:%.*]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP1:%.*]] = bitcast <4 x i32> [[TMP0]] to <16 x i8>
// CHECK-A32-SOFTFP-NEXT: [[VCVTFP2BF_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <4 x float>
// CHECK-A32-SOFTFP-NEXT: [[VCVTFP2BF1_I:%.*]] = call <4 x i16> @llvm.arm.neon.vcvtfp2bf.v4i16(<4 x float> [[VCVTFP2BF_I]])
// CHECK-A32-SOFTFP-NEXT: [[TMP2:%.*]] = bitcast <4 x i16> [[VCVTFP2BF1_I]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP3:%.*]] = bitcast <4 x bfloat> [[TMP2]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP4:%.*]] = bitcast <2 x i32> [[TMP3]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP5:%.*]] = bitcast <4 x bfloat> [[TMP4]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP6:%.*]] = bitcast <2 x i32> [[TMP5]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP7:%.*]] = bitcast <4 x bfloat> [[TMP6]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: ret <2 x i32> [[TMP7]]
//
bfloat16x4_t test_vcvt_bf16_f32(float32x4_t a) {
return vcvt_bf16_f32(a);
}
// CHECK-A64-LABEL: @test_vcvtq_low_bf16_f32(
// CHECK-A64-NEXT: entry:
// CHECK-A64-NEXT: [[TMP0:%.*]] = bitcast <4 x float> [[A:%.*]] to <4 x i32>
// CHECK-A64-NEXT: [[TMP1:%.*]] = bitcast <4 x i32> [[TMP0]] to <16 x i8>
// CHECK-A64-NEXT: [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <4 x float>
// CHECK-A64-NEXT: [[TMP3:%.*]] = fptrunc <4 x float> [[TMP2]] to <4 x bfloat>
// CHECK-A64-NEXT: [[TMP4:%.*]] = shufflevector <4 x bfloat> [[TMP3]], <4 x bfloat> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
// CHECK-A64-NEXT: ret <8 x bfloat> [[TMP4]]
//
// CHECK-A32-HARDFP-LABEL: @test_vcvtq_low_bf16_f32(
// CHECK-A32-HARDFP-NEXT: entry:
// CHECK-A32-HARDFP-NEXT: [[TMP0:%.*]] = bitcast i64 0 to <4 x bfloat>
// CHECK-A32-HARDFP-NEXT: [[TMP1:%.*]] = bitcast <4 x float> [[A:%.*]] to <4 x i32>
// CHECK-A32-HARDFP-NEXT: [[TMP2:%.*]] = bitcast <4 x i32> [[TMP1]] to <16 x i8>
// CHECK-A32-HARDFP-NEXT: [[VCVTFP2BF_I:%.*]] = bitcast <16 x i8> [[TMP2]] to <4 x float>
// CHECK-A32-HARDFP-NEXT: [[VCVTFP2BF1_I:%.*]] = call <4 x bfloat> @llvm.arm.neon.vcvtfp2bf.v4bf16(<4 x float> [[VCVTFP2BF_I]])
// CHECK-A32-HARDFP-NEXT: [[SHUFFLE_I:%.*]] = shufflevector <4 x bfloat> [[TMP0]], <4 x bfloat> [[VCVTFP2BF1_I]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
// CHECK-A32-HARDFP-NEXT: ret <8 x bfloat> [[SHUFFLE_I]]
//
// CHECK-A32-SOFTFP-LABEL: @test_vcvtq_low_bf16_f32(
// CHECK-A32-SOFTFP-NEXT: entry:
// CHECK-A32-SOFTFP-NEXT: [[TMP0:%.*]] = bitcast i64 0 to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP1:%.*]] = bitcast <4 x float> [[A:%.*]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP2:%.*]] = bitcast <4 x i32> [[TMP1]] to <16 x i8>
// CHECK-A32-SOFTFP-NEXT: [[VCVTFP2BF_I:%.*]] = bitcast <16 x i8> [[TMP2]] to <4 x float>
// CHECK-A32-SOFTFP-NEXT: [[VCVTFP2BF1_I:%.*]] = call <4 x i16> @llvm.arm.neon.vcvtfp2bf.v4i16(<4 x float> [[VCVTFP2BF_I]])
// CHECK-A32-SOFTFP-NEXT: [[TMP3:%.*]] = bitcast <4 x i16> [[VCVTFP2BF1_I]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP4:%.*]] = bitcast <4 x bfloat> [[TMP3]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP5:%.*]] = bitcast <2 x i32> [[TMP4]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP6:%.*]] = bitcast <4 x bfloat> [[TMP0]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP7:%.*]] = bitcast <4 x bfloat> [[TMP5]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP8:%.*]] = bitcast <2 x i32> [[TMP6]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP9:%.*]] = bitcast <2 x i32> [[TMP7]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[SHUFFLE_I:%.*]] = shufflevector <4 x bfloat> [[TMP8]], <4 x bfloat> [[TMP9]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
// CHECK-A32-SOFTFP-NEXT: [[TMP10:%.*]] = bitcast <8 x bfloat> [[SHUFFLE_I]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP11:%.*]] = bitcast <4 x i32> [[TMP10]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP12:%.*]] = bitcast <8 x bfloat> [[TMP11]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP13:%.*]] = bitcast <4 x i32> [[TMP12]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP14:%.*]] = bitcast <8 x bfloat> [[TMP13]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: ret <4 x i32> [[TMP14]]
//
bfloat16x8_t test_vcvtq_low_bf16_f32(float32x4_t a) {
return vcvtq_low_bf16_f32(a);
}
// CHECK-A64-LABEL: @test_vcvtq_high_bf16_f32(
// CHECK-A64-NEXT: entry:
// CHECK-A64-NEXT: [[TMP0:%.*]] = bitcast <8 x bfloat> [[INACTIVE:%.*]] to <8 x i16>
// CHECK-A64-NEXT: [[TMP1:%.*]] = bitcast <4 x float> [[A:%.*]] to <4 x i32>
// CHECK-A64-NEXT: [[TMP2:%.*]] = bitcast <8 x i16> [[TMP0]] to <16 x i8>
// CHECK-A64-NEXT: [[TMP3:%.*]] = bitcast <4 x i32> [[TMP1]] to <16 x i8>
// CHECK-A64-NEXT: [[TMP4:%.*]] = bitcast <16 x i8> [[TMP2]] to <8 x bfloat>
// CHECK-A64-NEXT: [[TMP5:%.*]] = shufflevector <8 x bfloat> [[TMP4]], <8 x bfloat> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
// CHECK-A64-NEXT: [[TMP6:%.*]] = bitcast <16 x i8> [[TMP3]] to <4 x float>
// CHECK-A64-NEXT: [[TMP7:%.*]] = fptrunc <4 x float> [[TMP6]] to <4 x bfloat>
// CHECK-A64-NEXT: [[TMP8:%.*]] = shufflevector <4 x bfloat> [[TMP5]], <4 x bfloat> [[TMP7]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
// CHECK-A64-NEXT: ret <8 x bfloat> [[TMP8]]
//
// CHECK-A32-HARDFP-LABEL: @test_vcvtq_high_bf16_f32(
// CHECK-A32-HARDFP-NEXT: entry:
// CHECK-A32-HARDFP-NEXT: [[TMP0:%.*]] = bitcast <4 x float> [[A:%.*]] to <4 x i32>
// CHECK-A32-HARDFP-NEXT: [[TMP1:%.*]] = bitcast <4 x i32> [[TMP0]] to <16 x i8>
// CHECK-A32-HARDFP-NEXT: [[VCVTFP2BF_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <4 x float>
// CHECK-A32-HARDFP-NEXT: [[VCVTFP2BF1_I:%.*]] = call <4 x bfloat> @llvm.arm.neon.vcvtfp2bf.v4bf16(<4 x float> [[VCVTFP2BF_I]])
// CHECK-A32-HARDFP-NEXT: [[SHUFFLE_I:%.*]] = shufflevector <8 x bfloat> [[INACTIVE:%.*]], <8 x bfloat> [[INACTIVE]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
// CHECK-A32-HARDFP-NEXT: [[SHUFFLE_I8:%.*]] = shufflevector <4 x bfloat> [[VCVTFP2BF1_I]], <4 x bfloat> [[SHUFFLE_I]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
// CHECK-A32-HARDFP-NEXT: ret <8 x bfloat> [[SHUFFLE_I8]]
//
// CHECK-A32-SOFTFP-LABEL: @test_vcvtq_high_bf16_f32(
// CHECK-A32-SOFTFP-NEXT: entry:
// CHECK-A32-SOFTFP-NEXT: [[TMP0:%.*]] = bitcast <4 x i32> [[INACTIVE_COERCE:%.*]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP1:%.*]] = bitcast <8 x bfloat> [[TMP0]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP2:%.*]] = bitcast <4 x i32> [[TMP1]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP3:%.*]] = bitcast <4 x float> [[A:%.*]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP4:%.*]] = bitcast <4 x i32> [[TMP3]] to <16 x i8>
// CHECK-A32-SOFTFP-NEXT: [[VCVTFP2BF_I:%.*]] = bitcast <16 x i8> [[TMP4]] to <4 x float>
// CHECK-A32-SOFTFP-NEXT: [[VCVTFP2BF1_I:%.*]] = call <4 x i16> @llvm.arm.neon.vcvtfp2bf.v4i16(<4 x float> [[VCVTFP2BF_I]])
// CHECK-A32-SOFTFP-NEXT: [[TMP5:%.*]] = bitcast <4 x i16> [[VCVTFP2BF1_I]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP6:%.*]] = bitcast <4 x bfloat> [[TMP5]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP7:%.*]] = bitcast <2 x i32> [[TMP6]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP8:%.*]] = bitcast <8 x bfloat> [[TMP2]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP9:%.*]] = bitcast <4 x i32> [[TMP8]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[SHUFFLE_I:%.*]] = shufflevector <8 x bfloat> [[TMP9]], <8 x bfloat> [[TMP9]], <4 x i32> <i32 0, i32 1, i32 2, i32 3>
// CHECK-A32-SOFTFP-NEXT: [[TMP10:%.*]] = bitcast <4 x bfloat> [[SHUFFLE_I]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP11:%.*]] = bitcast <2 x i32> [[TMP10]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP12:%.*]] = bitcast <4 x bfloat> [[TMP7]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP13:%.*]] = bitcast <4 x bfloat> [[TMP11]] to <2 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP14:%.*]] = bitcast <2 x i32> [[TMP12]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP15:%.*]] = bitcast <2 x i32> [[TMP13]] to <4 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[SHUFFLE_I17:%.*]] = shufflevector <4 x bfloat> [[TMP14]], <4 x bfloat> [[TMP15]], <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
// CHECK-A32-SOFTFP-NEXT: [[TMP16:%.*]] = bitcast <8 x bfloat> [[SHUFFLE_I17]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP17:%.*]] = bitcast <4 x i32> [[TMP16]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP18:%.*]] = bitcast <8 x bfloat> [[TMP17]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: [[TMP19:%.*]] = bitcast <4 x i32> [[TMP18]] to <8 x bfloat>
// CHECK-A32-SOFTFP-NEXT: [[TMP20:%.*]] = bitcast <8 x bfloat> [[TMP19]] to <4 x i32>
// CHECK-A32-SOFTFP-NEXT: ret <4 x i32> [[TMP20]]
//
bfloat16x8_t test_vcvtq_high_bf16_f32(bfloat16x8_t inactive, float32x4_t a) {
return vcvtq_high_bf16_f32(inactive, a);
}
// CHECK-A64-LABEL: @test_vcvth_bf16_f32(
// CHECK-A64-NEXT: entry:
// CHECK-A64-NEXT: [[TMP0:%.*]] = fptrunc float [[A:%.*]] to bfloat
// CHECK-A64-NEXT: ret bfloat [[TMP0]]
//
// CHECK-A32-HARDFP-LABEL: @test_vcvth_bf16_f32(
// CHECK-A32-HARDFP-NEXT: entry:
// CHECK-A32-HARDFP-NEXT: [[VCVTBFP2BF_I:%.*]] = call bfloat @llvm.arm.neon.vcvtbfp2bf(float [[A:%.*]])
// CHECK-A32-HARDFP-NEXT: ret bfloat [[VCVTBFP2BF_I]]
//
// CHECK-A32-SOFTFP-LABEL: @test_vcvth_bf16_f32(
// CHECK-A32-SOFTFP-NEXT: entry:
// CHECK-A32-SOFTFP-NEXT: [[VCVTBFP2BF_I:%.*]] = call bfloat @llvm.arm.neon.vcvtbfp2bf(float [[A:%.*]])
// CHECK-A32-SOFTFP-NEXT: ret bfloat [[VCVTBFP2BF_I]]
//
bfloat16_t test_vcvth_bf16_f32(float32_t a) {
return vcvth_bf16_f32(a);
}
// CHECK-LABEL: @test_vcvtah_f32_bf16(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = bitcast bfloat [[A:%.*]] to i16
// CHECK-NEXT: [[CONV_I:%.*]] = zext i16 [[TMP0]] to i32
// CHECK-NEXT: [[SHL_I:%.*]] = shl i32 [[CONV_I]], 16
// CHECK-NEXT: [[TMP1:%.*]] = bitcast i32 [[SHL_I]] to float
// CHECK-NEXT: ret float [[TMP1]]
//
float32_t test_vcvtah_f32_bf16(bfloat16_t a) {
return vcvtah_f32_bf16(a);
}
|
