1 /* Copyright (C) 2003-2013 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 Intel C++ Compiler 25 User Guide and Reference, version 9.0. */ 26 27 #ifndef _PMMINTRIN_H_INCLUDED 28 #define _PMMINTRIN_H_INCLUDED 29 30 #ifndef __SSE3__ 31 # error "SSE3 instruction set not enabled" 32 #else 33 34 /* We need definitions from the SSE2 and SSE header files*/ 35 #include <emmintrin.h> 36 37 /* Additional bits in the MXCSR. */ 38 #define _MM_DENORMALS_ZERO_MASK 0x0040 39 #define _MM_DENORMALS_ZERO_ON 0x0040 40 #define _MM_DENORMALS_ZERO_OFF 0x0000 41 42 #define _MM_SET_DENORMALS_ZERO_MODE(mode) \ 43 _mm_setcsr ((_mm_getcsr () & ~_MM_DENORMALS_ZERO_MASK) | (mode)) 44 #define _MM_GET_DENORMALS_ZERO_MODE() \ 45 (_mm_getcsr() & _MM_DENORMALS_ZERO_MASK) 46 47 extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 48 _mm_addsub_ps (__m128 __X, __m128 __Y) 49 { 50 return (__m128) __builtin_ia32_addsubps ((__v4sf)__X, (__v4sf)__Y); 51 } 52 53 extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 54 _mm_hadd_ps (__m128 __X, __m128 __Y) 55 { 56 return (__m128) __builtin_ia32_haddps ((__v4sf)__X, (__v4sf)__Y); 57 } 58 59 extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 60 _mm_hsub_ps (__m128 __X, __m128 __Y) 61 { 62 return (__m128) __builtin_ia32_hsubps ((__v4sf)__X, (__v4sf)__Y); 63 } 64 65 extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 66 _mm_movehdup_ps (__m128 __X) 67 { 68 return (__m128) __builtin_ia32_movshdup ((__v4sf)__X); 69 } 70 71 extern __inline __m128 __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 72 _mm_moveldup_ps (__m128 __X) 73 { 74 return (__m128) __builtin_ia32_movsldup ((__v4sf)__X); 75 } 76 77 extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 78 _mm_addsub_pd (__m128d __X, __m128d __Y) 79 { 80 return (__m128d) __builtin_ia32_addsubpd ((__v2df)__X, (__v2df)__Y); 81 } 82 83 extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 84 _mm_hadd_pd (__m128d __X, __m128d __Y) 85 { 86 return (__m128d) __builtin_ia32_haddpd ((__v2df)__X, (__v2df)__Y); 87 } 88 89 extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 90 _mm_hsub_pd (__m128d __X, __m128d __Y) 91 { 92 return (__m128d) __builtin_ia32_hsubpd ((__v2df)__X, (__v2df)__Y); 93 } 94 95 extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 96 _mm_loaddup_pd (double const *__P) 97 { 98 return _mm_load1_pd (__P); 99 } 100 101 extern __inline __m128d __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 102 _mm_movedup_pd (__m128d __X) 103 { 104 return _mm_shuffle_pd (__X, __X, _MM_SHUFFLE2 (0,0)); 105 } 106 107 extern __inline __m128i __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 108 _mm_lddqu_si128 (__m128i const *__P) 109 { 110 return (__m128i) __builtin_ia32_lddqu ((char const *)__P); 111 } 112 113 extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 114 _mm_monitor (void const * __P, unsigned int __E, unsigned int __H) 115 { 116 __builtin_ia32_monitor (__P, __E, __H); 117 } 118 119 extern __inline void __attribute__((__gnu_inline__, __always_inline__, __artificial__)) 120 _mm_mwait (unsigned int __E, unsigned int __H) 121 { 122 __builtin_ia32_mwait (__E, __H); 123 } 124 125 #endif /* __SSE3__ */ 126 127 #endif /* _PMMINTRIN_H_INCLUDED */ 128