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
|
/*===---- avx512erintrin.h - AVX512ER intrinsics ---------------------------===
*
* Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
* See https://llvm.org/LICENSE.txt for license information.
* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
*
*===-----------------------------------------------------------------------===
*/
#ifndef __IMMINTRIN_H
#error "Never use <avx512erintrin.h> directly; include <immintrin.h> instead."
#endif
#ifndef __AVX512ERINTRIN_H
#define __AVX512ERINTRIN_H
/* exp2a23 */
#define _mm512_exp2a23_round_pd(A, R) \
((__m512d)__builtin_ia32_exp2pd_mask((__v8df)(__m512d)(A), \
(__v8df)_mm512_setzero_pd(), \
(__mmask8)-1, (int)(R)))
#define _mm512_mask_exp2a23_round_pd(S, M, A, R) \
((__m512d)__builtin_ia32_exp2pd_mask((__v8df)(__m512d)(A), \
(__v8df)(__m512d)(S), (__mmask8)(M), \
(int)(R)))
#define _mm512_maskz_exp2a23_round_pd(M, A, R) \
((__m512d)__builtin_ia32_exp2pd_mask((__v8df)(__m512d)(A), \
(__v8df)_mm512_setzero_pd(), \
(__mmask8)(M), (int)(R)))
#define _mm512_exp2a23_pd(A) \
_mm512_exp2a23_round_pd((A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_mask_exp2a23_pd(S, M, A) \
_mm512_mask_exp2a23_round_pd((S), (M), (A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_maskz_exp2a23_pd(M, A) \
_mm512_maskz_exp2a23_round_pd((M), (A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_exp2a23_round_ps(A, R) \
((__m512)__builtin_ia32_exp2ps_mask((__v16sf)(__m512)(A), \
(__v16sf)_mm512_setzero_ps(), \
(__mmask16)-1, (int)(R)))
#define _mm512_mask_exp2a23_round_ps(S, M, A, R) \
((__m512)__builtin_ia32_exp2ps_mask((__v16sf)(__m512)(A), \
(__v16sf)(__m512)(S), (__mmask16)(M), \
(int)(R)))
#define _mm512_maskz_exp2a23_round_ps(M, A, R) \
((__m512)__builtin_ia32_exp2ps_mask((__v16sf)(__m512)(A), \
(__v16sf)_mm512_setzero_ps(), \
(__mmask16)(M), (int)(R)))
#define _mm512_exp2a23_ps(A) \
_mm512_exp2a23_round_ps((A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_mask_exp2a23_ps(S, M, A) \
_mm512_mask_exp2a23_round_ps((S), (M), (A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_maskz_exp2a23_ps(M, A) \
_mm512_maskz_exp2a23_round_ps((M), (A), _MM_FROUND_CUR_DIRECTION)
/* rsqrt28 */
#define _mm512_rsqrt28_round_pd(A, R) \
((__m512d)__builtin_ia32_rsqrt28pd_mask((__v8df)(__m512d)(A), \
(__v8df)_mm512_setzero_pd(), \
(__mmask8)-1, (int)(R)))
#define _mm512_mask_rsqrt28_round_pd(S, M, A, R) \
((__m512d)__builtin_ia32_rsqrt28pd_mask((__v8df)(__m512d)(A), \
(__v8df)(__m512d)(S), (__mmask8)(M), \
(int)(R)))
#define _mm512_maskz_rsqrt28_round_pd(M, A, R) \
((__m512d)__builtin_ia32_rsqrt28pd_mask((__v8df)(__m512d)(A), \
(__v8df)_mm512_setzero_pd(), \
(__mmask8)(M), (int)(R)))
#define _mm512_rsqrt28_pd(A) \
_mm512_rsqrt28_round_pd((A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_mask_rsqrt28_pd(S, M, A) \
_mm512_mask_rsqrt28_round_pd((S), (M), (A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_maskz_rsqrt28_pd(M, A) \
_mm512_maskz_rsqrt28_round_pd((M), (A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_rsqrt28_round_ps(A, R) \
((__m512)__builtin_ia32_rsqrt28ps_mask((__v16sf)(__m512)(A), \
(__v16sf)_mm512_setzero_ps(), \
(__mmask16)-1, (int)(R)))
#define _mm512_mask_rsqrt28_round_ps(S, M, A, R) \
((__m512)__builtin_ia32_rsqrt28ps_mask((__v16sf)(__m512)(A), \
(__v16sf)(__m512)(S), (__mmask16)(M), \
(int)(R)))
#define _mm512_maskz_rsqrt28_round_ps(M, A, R) \
((__m512)__builtin_ia32_rsqrt28ps_mask((__v16sf)(__m512)(A), \
(__v16sf)_mm512_setzero_ps(), \
(__mmask16)(M), (int)(R)))
#define _mm512_rsqrt28_ps(A) \
_mm512_rsqrt28_round_ps((A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_mask_rsqrt28_ps(S, M, A) \
_mm512_mask_rsqrt28_round_ps((S), (M), A, _MM_FROUND_CUR_DIRECTION)
#define _mm512_maskz_rsqrt28_ps(M, A) \
_mm512_maskz_rsqrt28_round_ps((M), (A), _MM_FROUND_CUR_DIRECTION)
#define _mm_rsqrt28_round_ss(A, B, R) \
((__m128)__builtin_ia32_rsqrt28ss_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(B), \
(__v4sf)_mm_setzero_ps(), \
(__mmask8)-1, (int)(R)))
#define _mm_mask_rsqrt28_round_ss(S, M, A, B, R) \
((__m128)__builtin_ia32_rsqrt28ss_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(B), \
(__v4sf)(__m128)(S), \
(__mmask8)(M), (int)(R)))
#define _mm_maskz_rsqrt28_round_ss(M, A, B, R) \
((__m128)__builtin_ia32_rsqrt28ss_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(B), \
(__v4sf)_mm_setzero_ps(), \
(__mmask8)(M), (int)(R)))
#define _mm_rsqrt28_ss(A, B) \
_mm_rsqrt28_round_ss((A), (B), _MM_FROUND_CUR_DIRECTION)
#define _mm_mask_rsqrt28_ss(S, M, A, B) \
_mm_mask_rsqrt28_round_ss((S), (M), (A), (B), _MM_FROUND_CUR_DIRECTION)
#define _mm_maskz_rsqrt28_ss(M, A, B) \
_mm_maskz_rsqrt28_round_ss((M), (A), (B), _MM_FROUND_CUR_DIRECTION)
#define _mm_rsqrt28_round_sd(A, B, R) \
((__m128d)__builtin_ia32_rsqrt28sd_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(B), \
(__v2df)_mm_setzero_pd(), \
(__mmask8)-1, (int)(R)))
#define _mm_mask_rsqrt28_round_sd(S, M, A, B, R) \
((__m128d)__builtin_ia32_rsqrt28sd_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(B), \
(__v2df)(__m128d)(S), \
(__mmask8)(M), (int)(R)))
#define _mm_maskz_rsqrt28_round_sd(M, A, B, R) \
((__m128d)__builtin_ia32_rsqrt28sd_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(B), \
(__v2df)_mm_setzero_pd(), \
(__mmask8)(M), (int)(R)))
#define _mm_rsqrt28_sd(A, B) \
_mm_rsqrt28_round_sd((A), (B), _MM_FROUND_CUR_DIRECTION)
#define _mm_mask_rsqrt28_sd(S, M, A, B) \
_mm_mask_rsqrt28_round_sd((S), (M), (A), (B), _MM_FROUND_CUR_DIRECTION)
#define _mm_maskz_rsqrt28_sd(M, A, B) \
_mm_maskz_rsqrt28_round_sd((M), (A), (B), _MM_FROUND_CUR_DIRECTION)
/* rcp28 */
#define _mm512_rcp28_round_pd(A, R) \
((__m512d)__builtin_ia32_rcp28pd_mask((__v8df)(__m512d)(A), \
(__v8df)_mm512_setzero_pd(), \
(__mmask8)-1, (int)(R)))
#define _mm512_mask_rcp28_round_pd(S, M, A, R) \
((__m512d)__builtin_ia32_rcp28pd_mask((__v8df)(__m512d)(A), \
(__v8df)(__m512d)(S), (__mmask8)(M), \
(int)(R)))
#define _mm512_maskz_rcp28_round_pd(M, A, R) \
((__m512d)__builtin_ia32_rcp28pd_mask((__v8df)(__m512d)(A), \
(__v8df)_mm512_setzero_pd(), \
(__mmask8)(M), (int)(R)))
#define _mm512_rcp28_pd(A) \
_mm512_rcp28_round_pd((A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_mask_rcp28_pd(S, M, A) \
_mm512_mask_rcp28_round_pd((S), (M), (A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_maskz_rcp28_pd(M, A) \
_mm512_maskz_rcp28_round_pd((M), (A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_rcp28_round_ps(A, R) \
((__m512)__builtin_ia32_rcp28ps_mask((__v16sf)(__m512)(A), \
(__v16sf)_mm512_setzero_ps(), \
(__mmask16)-1, (int)(R)))
#define _mm512_mask_rcp28_round_ps(S, M, A, R) \
((__m512)__builtin_ia32_rcp28ps_mask((__v16sf)(__m512)(A), \
(__v16sf)(__m512)(S), (__mmask16)(M), \
(int)(R)))
#define _mm512_maskz_rcp28_round_ps(M, A, R) \
((__m512)__builtin_ia32_rcp28ps_mask((__v16sf)(__m512)(A), \
(__v16sf)_mm512_setzero_ps(), \
(__mmask16)(M), (int)(R)))
#define _mm512_rcp28_ps(A) \
_mm512_rcp28_round_ps((A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_mask_rcp28_ps(S, M, A) \
_mm512_mask_rcp28_round_ps((S), (M), (A), _MM_FROUND_CUR_DIRECTION)
#define _mm512_maskz_rcp28_ps(M, A) \
_mm512_maskz_rcp28_round_ps((M), (A), _MM_FROUND_CUR_DIRECTION)
#define _mm_rcp28_round_ss(A, B, R) \
((__m128)__builtin_ia32_rcp28ss_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(B), \
(__v4sf)_mm_setzero_ps(), \
(__mmask8)-1, (int)(R)))
#define _mm_mask_rcp28_round_ss(S, M, A, B, R) \
((__m128)__builtin_ia32_rcp28ss_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(B), \
(__v4sf)(__m128)(S), \
(__mmask8)(M), (int)(R)))
#define _mm_maskz_rcp28_round_ss(M, A, B, R) \
((__m128)__builtin_ia32_rcp28ss_round_mask((__v4sf)(__m128)(A), \
(__v4sf)(__m128)(B), \
(__v4sf)_mm_setzero_ps(), \
(__mmask8)(M), (int)(R)))
#define _mm_rcp28_ss(A, B) \
_mm_rcp28_round_ss((A), (B), _MM_FROUND_CUR_DIRECTION)
#define _mm_mask_rcp28_ss(S, M, A, B) \
_mm_mask_rcp28_round_ss((S), (M), (A), (B), _MM_FROUND_CUR_DIRECTION)
#define _mm_maskz_rcp28_ss(M, A, B) \
_mm_maskz_rcp28_round_ss((M), (A), (B), _MM_FROUND_CUR_DIRECTION)
#define _mm_rcp28_round_sd(A, B, R) \
((__m128d)__builtin_ia32_rcp28sd_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(B), \
(__v2df)_mm_setzero_pd(), \
(__mmask8)-1, (int)(R)))
#define _mm_mask_rcp28_round_sd(S, M, A, B, R) \
((__m128d)__builtin_ia32_rcp28sd_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(B), \
(__v2df)(__m128d)(S), \
(__mmask8)(M), (int)(R)))
#define _mm_maskz_rcp28_round_sd(M, A, B, R) \
((__m128d)__builtin_ia32_rcp28sd_round_mask((__v2df)(__m128d)(A), \
(__v2df)(__m128d)(B), \
(__v2df)_mm_setzero_pd(), \
(__mmask8)(M), (int)(R)))
#define _mm_rcp28_sd(A, B) \
_mm_rcp28_round_sd((A), (B), _MM_FROUND_CUR_DIRECTION)
#define _mm_mask_rcp28_sd(S, M, A, B) \
_mm_mask_rcp28_round_sd((S), (M), (A), (B), _MM_FROUND_CUR_DIRECTION)
#define _mm_maskz_rcp28_sd(M, A, B) \
_mm_maskz_rcp28_round_sd((M), (A), (B), _MM_FROUND_CUR_DIRECTION)
#endif /* __AVX512ERINTRIN_H */
|
