| 1 | /* Copyright (C) 2007-2015 Free Software Foundation, Inc. | 
 
 
 
 
 | 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 | /* Implemented from the specification included in the AMD Programmers | 
 
 
 
 
 | 25 | Manual Update, version 2.x */ | 
 
 
 
 
 | 26 |  | 
 
 
 
 
 | 27 | #ifndef _AMMINTRIN_H_INCLUDED | 
 
 
 
 
 | 28 | #define _AMMINTRIN_H_INCLUDED | 
 
 
 
 
 | 29 |  | 
 
 
 
 
 | 30 | /* We need definitions from the SSE3, SSE2 and SSE header files*/ | 
 
 
 
 
 | 31 | #include <pmmintrin.h> | 
 
 
 
 
 | 32 |  | 
 
 
 
 
 | 33 | #ifndef __SSE4A__ | 
 
 
 
 
 | 34 | #pragma GCC push_options | 
 
 
 
 
 | 35 | #pragma GCC target("sse4a") | 
 
 
 
 
 | 36 | #define __DISABLE_SSE4A__ | 
 
 
 
 
 | 37 | #endif /* __SSE4A__ */ | 
 
 
 
 
 | 38 |  | 
 
 
 
 
 | 39 | extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | 
 
 
 
 
 | 40 | _mm_stream_sd (double * __P, __m128d __Y) | 
 
 
 
 
 | 41 | { | 
 
 
 
 
 | 42 | __builtin_ia32_movntsd (__P, (__v2df) __Y); | 
 
 
 
 
 | 43 | } | 
 
 
 
 
 | 44 |  | 
 
 
 
 
 | 45 | extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | 
 
 
 
 
 | 46 | _mm_stream_ss (float * __P, __m128 __Y) | 
 
 
 
 
 | 47 | { | 
 
 
 
 
 | 48 | __builtin_ia32_movntss (__P, (__v4sf) __Y); | 
 
 
 
 
 | 49 | } | 
 
 
 
 
 | 50 |  | 
 
 
 
 
 | 51 | extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | 
 
 
 
 
 | 52 | _mm_extract_si64 (__m128i __X, __m128i __Y) | 
 
 
 
 
 | 53 | { | 
 
 
 
 
 | 54 | return (__m128i) __builtin_ia32_extrq ((__v2di) __X, (__v16qi) __Y); | 
 
 
 
 
 | 55 | } | 
 
 
 
 
 | 56 |  | 
 
 
 
 
 | 57 | #ifdef __OPTIMIZE__ | 
 
 
 
 
 | 58 | extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | 
 
 
 
 
 | 59 | _mm_extracti_si64 (__m128i __X, unsigned const int __I, unsigned const int __L) | 
 
 
 
 
 | 60 | { | 
 
 
 
 
 | 61 | return (__m128i) __builtin_ia32_extrqi ((__v2di) __X, __I, __L); | 
 
 
 
 
 | 62 | } | 
 
 
 
 
 | 63 | #else | 
 
 
 
 
 | 64 | #define _mm_extracti_si64(X, I, L)                                      \ | 
 
 
 
 
 | 65 | ((__m128i) __builtin_ia32_extrqi ((__v2di)(__m128i)(X),               \ | 
 
 
 
 
 | 66 | (unsigned int)(I), (unsigned int)(L))) | 
 
 
 
 
 | 67 | #endif | 
 
 
 
 
 | 68 |  | 
 
 
 
 
 | 69 | extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | 
 
 
 
 
 | 70 | _mm_insert_si64 (__m128i __X,__m128i __Y) | 
 
 
 
 
 | 71 | { | 
 
 
 
 
 | 72 | return (__m128i) __builtin_ia32_insertq ((__v2di)__X, (__v2di)__Y); | 
 
 
 
 
 | 73 | } | 
 
 
 
 
 | 74 |  | 
 
 
 
 
 | 75 | #ifdef __OPTIMIZE__ | 
 
 
 
 
 | 76 | extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__)) | 
 
 
 
 
 | 77 | _mm_inserti_si64(__m128i __X, __m128i __Y, unsigned const int __I, unsigned const int __L) | 
 
 
 
 
 | 78 | { | 
 
 
 
 
 | 79 | return (__m128i) __builtin_ia32_insertqi ((__v2di)__X, (__v2di)__Y, __I, __L); | 
 
 
 
 
 | 80 | } | 
 
 
 
 
 | 81 | #else | 
 
 
 
 
 | 82 | #define _mm_inserti_si64(X, Y, I, L)                                    \ | 
 
 
 
 
 | 83 | ((__m128i) __builtin_ia32_insertqi ((__v2di)(__m128i)(X),             \ | 
 
 
 
 
 | 84 | (__v2di)(__m128i)(Y),             \ | 
 
 
 
 
 | 85 | (unsigned int)(I), (unsigned int)(L))) | 
 
 
 
 
 | 86 | #endif | 
 
 
 
 
 | 87 |  | 
 
 
 
 
 | 88 | #ifdef __DISABLE_SSE4A__ | 
 
 
 
 
 | 89 | #undef __DISABLE_SSE4A__ | 
 
 
 
 
 | 90 | #pragma GCC pop_options | 
 
 
 
 
 | 91 | #endif /* __DISABLE_SSE4A__ */ | 
 
 
 
 
 | 92 |  | 
 
 
 
 
 | 93 | #endif /* _AMMINTRIN_H_INCLUDED */ |