]>
Commit | Line | Data |
---|---|---|
83ffe9cd | 1 | /* Copyright (C) 2015-2023 Free Software Foundation, Inc. |
5fbb13a7 KY |
2 | |
3 | This file is part of GCC. | |
4 | ||
5 | GCC is free software; you can redistribute it and/or modify | |
6 | it under the terms of the GNU General Public License as published by | |
7 | the Free Software Foundation; either version 3, or (at your option) | |
8 | any later version. | |
9 | ||
10 | GCC is distributed in the hope that it will be useful, | |
11 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
12 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
13 | GNU General Public License for more details. | |
14 | ||
15 | Under Section 7 of GPL version 3, you are granted additional | |
16 | permissions described in the GCC Runtime Library Exception, version | |
17 | 3.1, as published by the Free Software Foundation. | |
18 | ||
19 | You should have received a copy of the GNU General Public License and | |
20 | a copy of the GCC Runtime Library Exception along with this program; | |
21 | see the files COPYING3 and COPYING.RUNTIME respectively. If not, see | |
22 | <http://www.gnu.org/licenses/>. */ | |
23 | ||
24 | #if !defined _IMMINTRIN_H_INCLUDED | |
25 | # error "Never use <avx5124fmapsintrin.h> directly; include <x86intrin.h> instead." | |
26 | #endif | |
27 | ||
28 | #ifndef _AVX5124FMAPSINTRIN_H_INCLUDED | |
29 | #define _AVX5124FMAPSINTRIN_H_INCLUDED | |
30 | ||
31 | #ifndef __AVX5124FMAPS__ | |
32 | #pragma GCC push_options | |
2b77bcbb | 33 | #pragma GCC target("avx5124fmaps,evex512") |
5fbb13a7 KY |
34 | #define __DISABLE_AVX5124FMAPS__ |
35 | #endif /* __AVX5124FMAPS__ */ | |
36 | ||
37 | extern __inline __m512 | |
38 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
39 | _mm512_4fmadd_ps (__m512 __A, __m512 __B, __m512 __C, | |
40 | __m512 __D, __m512 __E, __m128 *__F) | |
41 | { | |
42 | return (__m512) __builtin_ia32_4fmaddps ((__v16sf) __B, | |
43 | (__v16sf) __C, | |
44 | (__v16sf) __D, | |
45 | (__v16sf) __E, | |
46 | (__v16sf) __A, | |
47 | (const __v4sf *) __F); | |
48 | } | |
49 | ||
50 | extern __inline __m512 | |
51 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
52 | _mm512_mask_4fmadd_ps (__m512 __A, __mmask16 __U, __m512 __B, | |
53 | __m512 __C, __m512 __D, __m512 __E, __m128 *__F) | |
54 | { | |
55 | return (__m512) __builtin_ia32_4fmaddps_mask ((__v16sf) __B, | |
56 | (__v16sf) __C, | |
57 | (__v16sf) __D, | |
58 | (__v16sf) __E, | |
59 | (__v16sf) __A, | |
60 | (const __v4sf *) __F, | |
61 | (__v16sf) __A, | |
62 | (__mmask16) __U); | |
63 | } | |
64 | ||
65 | extern __inline __m512 | |
66 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
67 | _mm512_maskz_4fmadd_ps (__mmask16 __U, | |
68 | __m512 __A, __m512 __B, __m512 __C, | |
69 | __m512 __D, __m512 __E, __m128 *__F) | |
70 | { | |
71 | return (__m512) __builtin_ia32_4fmaddps_mask ((__v16sf) __B, | |
72 | (__v16sf) __C, | |
73 | (__v16sf) __D, | |
74 | (__v16sf) __E, | |
75 | (__v16sf) __A, | |
76 | (const __v4sf *) __F, | |
77 | (__v16sf) _mm512_setzero_ps (), | |
78 | (__mmask16) __U); | |
79 | } | |
80 | ||
81 | extern __inline __m128 | |
82 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
83 | _mm_4fmadd_ss (__m128 __A, __m128 __B, __m128 __C, | |
84 | __m128 __D, __m128 __E, __m128 *__F) | |
85 | { | |
86 | return (__m128) __builtin_ia32_4fmaddss ((__v4sf) __B, | |
87 | (__v4sf) __C, | |
88 | (__v4sf) __D, | |
89 | (__v4sf) __E, | |
90 | (__v4sf) __A, | |
91 | (const __v4sf *) __F); | |
92 | } | |
93 | ||
94 | extern __inline __m128 | |
95 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
96 | _mm_mask_4fmadd_ss (__m128 __A, __mmask8 __U, __m128 __B, __m128 __C, | |
97 | __m128 __D, __m128 __E, __m128 *__F) | |
98 | { | |
99 | return (__m128) __builtin_ia32_4fmaddss_mask ((__v4sf) __B, | |
100 | (__v4sf) __C, | |
101 | (__v4sf) __D, | |
102 | (__v4sf) __E, | |
103 | (__v4sf) __A, | |
104 | (const __v4sf *) __F, | |
105 | (__v4sf) __A, | |
106 | (__mmask8) __U); | |
107 | } | |
108 | ||
109 | extern __inline __m128 | |
110 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
111 | _mm_maskz_4fmadd_ss (__mmask8 __U, __m128 __A, __m128 __B, __m128 __C, | |
112 | __m128 __D, __m128 __E, __m128 *__F) | |
113 | { | |
114 | return (__m128) __builtin_ia32_4fmaddss_mask ((__v4sf) __B, | |
115 | (__v4sf) __C, | |
116 | (__v4sf) __D, | |
117 | (__v4sf) __E, | |
118 | (__v4sf) __A, | |
119 | (const __v4sf *) __F, | |
120 | (__v4sf) _mm_setzero_ps (), | |
121 | (__mmask8) __U); | |
122 | } | |
123 | ||
124 | extern __inline __m512 | |
125 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
126 | _mm512_4fnmadd_ps (__m512 __A, __m512 __B, __m512 __C, | |
127 | __m512 __D, __m512 __E, __m128 *__F) | |
128 | { | |
129 | return (__m512) __builtin_ia32_4fnmaddps ((__v16sf) __B, | |
130 | (__v16sf) __C, | |
131 | (__v16sf) __D, | |
132 | (__v16sf) __E, | |
133 | (__v16sf) __A, | |
134 | (const __v4sf *) __F); | |
135 | } | |
136 | ||
137 | extern __inline __m512 | |
138 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
139 | _mm512_mask_4fnmadd_ps (__m512 __A, __mmask16 __U, __m512 __B, | |
140 | __m512 __C, __m512 __D, __m512 __E, __m128 *__F) | |
141 | { | |
142 | return (__m512) __builtin_ia32_4fnmaddps_mask ((__v16sf) __B, | |
143 | (__v16sf) __C, | |
144 | (__v16sf) __D, | |
145 | (__v16sf) __E, | |
146 | (__v16sf) __A, | |
147 | (const __v4sf *) __F, | |
148 | (__v16sf) __A, | |
149 | (__mmask16) __U); | |
150 | } | |
151 | ||
152 | extern __inline __m512 | |
153 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
154 | _mm512_maskz_4fnmadd_ps (__mmask16 __U, | |
155 | __m512 __A, __m512 __B, __m512 __C, | |
156 | __m512 __D, __m512 __E, __m128 *__F) | |
157 | { | |
158 | return (__m512) __builtin_ia32_4fnmaddps_mask ((__v16sf) __B, | |
159 | (__v16sf) __C, | |
160 | (__v16sf) __D, | |
161 | (__v16sf) __E, | |
162 | (__v16sf) __A, | |
163 | (const __v4sf *) __F, | |
164 | (__v16sf) _mm512_setzero_ps (), | |
165 | (__mmask16) __U); | |
166 | } | |
167 | ||
168 | extern __inline __m128 | |
169 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
170 | _mm_4fnmadd_ss (__m128 __A, __m128 __B, __m128 __C, | |
171 | __m128 __D, __m128 __E, __m128 *__F) | |
172 | { | |
173 | return (__m128) __builtin_ia32_4fnmaddss ((__v4sf) __B, | |
174 | (__v4sf) __C, | |
175 | (__v4sf) __D, | |
176 | (__v4sf) __E, | |
177 | (__v4sf) __A, | |
178 | (const __v4sf *) __F); | |
179 | } | |
180 | ||
181 | extern __inline __m128 | |
182 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
183 | _mm_mask_4fnmadd_ss (__m128 __A, __mmask8 __U, __m128 __B, __m128 __C, | |
184 | __m128 __D, __m128 __E, __m128 *__F) | |
185 | { | |
186 | return (__m128) __builtin_ia32_4fnmaddss_mask ((__v4sf) __B, | |
187 | (__v4sf) __C, | |
188 | (__v4sf) __D, | |
189 | (__v4sf) __E, | |
190 | (__v4sf) __A, | |
191 | (const __v4sf *) __F, | |
192 | (__v4sf) __A, | |
193 | (__mmask8) __U); | |
194 | } | |
195 | ||
196 | extern __inline __m128 | |
197 | __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | |
198 | _mm_maskz_4fnmadd_ss (__mmask8 __U, __m128 __A, __m128 __B, __m128 __C, | |
199 | __m128 __D, __m128 __E, __m128 *__F) | |
200 | { | |
201 | return (__m128) __builtin_ia32_4fnmaddss_mask ((__v4sf) __B, | |
202 | (__v4sf) __C, | |
203 | (__v4sf) __D, | |
204 | (__v4sf) __E, | |
205 | (__v4sf) __A, | |
206 | (const __v4sf *) __F, | |
207 | (__v4sf) _mm_setzero_ps (), | |
208 | (__mmask8) __U); | |
209 | } | |
210 | ||
211 | #ifdef __DISABLE_AVX5124FMAPS__ | |
212 | #undef __DISABLE_AVX5124FMAPS__ | |
213 | #pragma GCC pop_options | |
214 | #endif /* __DISABLE_AVX5124FMAPS__ */ | |
215 | ||
216 | #endif /* _AVX5124FMAPSINTRIN_H_INCLUDED */ |