cpu.h (5385B)
1 /* 2 * Copyright (c) 2000, 2001, 2002 Fabrice Bellard 3 * 4 * This file is part of FFmpeg. 5 * 6 * FFmpeg is free software; you can redistribute it and/or 7 * modify it under the terms of the GNU Lesser General Public 8 * License as published by the Free Software Foundation; either 9 * version 2.1 of the License, or (at your option) any later version. 10 * 11 * FFmpeg is distributed in the hope that it will be useful, 12 * but WITHOUT ANY WARRANTY; without even the implied warranty of 13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 14 * Lesser General Public License for more details. 15 * 16 * You should have received a copy of the GNU Lesser General Public 17 * License along with FFmpeg; if not, write to the Free Software 18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA 19 */ 20 21 #ifndef AVUTIL_CPU_H 22 #define AVUTIL_CPU_H 23 24 #include <stddef.h> 25 26 #define AV_CPU_FLAG_FORCE 0x80000000 /* force usage of selected flags (OR) */ 27 28 /* lower 16 bits - CPU features */ 29 #define AV_CPU_FLAG_MMX 0x0001 ///< standard MMX 30 #define AV_CPU_FLAG_MMXEXT 0x0002 ///< SSE integer functions or AMD MMX ext 31 #define AV_CPU_FLAG_MMX2 0x0002 ///< SSE integer functions or AMD MMX ext 32 #define AV_CPU_FLAG_3DNOW 0x0004 ///< AMD 3DNOW 33 #define AV_CPU_FLAG_SSE 0x0008 ///< SSE functions 34 #define AV_CPU_FLAG_SSE2 0x0010 ///< PIV SSE2 functions 35 #define AV_CPU_FLAG_SSE2SLOW \ 36 0x40000000 ///< SSE2 supported, but usually not faster 37 ///< than regular MMX/SSE (e.g. Core1) 38 #define AV_CPU_FLAG_3DNOWEXT 0x0020 ///< AMD 3DNowExt 39 #define AV_CPU_FLAG_SSE3 0x0040 ///< Prescott SSE3 functions 40 #define AV_CPU_FLAG_SSE3SLOW \ 41 0x20000000 ///< SSE3 supported, but usually not faster 42 ///< than regular MMX/SSE (e.g. Core1) 43 #define AV_CPU_FLAG_SSSE3 0x0080 ///< Conroe SSSE3 functions 44 #define AV_CPU_FLAG_SSSE3SLOW \ 45 0x4000000 ///< SSSE3 supported, but usually not faster 46 #define AV_CPU_FLAG_ATOM \ 47 0x10000000 ///< Atom processor, some SSSE3 instructions are slower 48 #define AV_CPU_FLAG_SSE4 0x0100 ///< Penryn SSE4.1 functions 49 #define AV_CPU_FLAG_SSE42 0x0200 ///< Nehalem SSE4.2 functions 50 #define AV_CPU_FLAG_AESNI 0x80000 ///< Advanced Encryption Standard functions 51 #define AV_CPU_FLAG_AVX \ 52 0x4000 ///< AVX functions: requires OS support even if YMM registers aren't 53 ///< used 54 #define AV_CPU_FLAG_AVXSLOW \ 55 0x8000000 ///< AVX supported, but slow when using YMM registers (e.g. 56 ///< Bulldozer) 57 #define AV_CPU_FLAG_XOP 0x0400 ///< Bulldozer XOP functions 58 #define AV_CPU_FLAG_FMA4 0x0800 ///< Bulldozer FMA4 functions 59 #define AV_CPU_FLAG_CMOV 0x1000 ///< supports cmov instruction 60 #define AV_CPU_FLAG_AVX2 \ 61 0x8000 ///< AVX2 functions: requires OS support even if YMM registers aren't 62 ///< used 63 #define AV_CPU_FLAG_FMA3 0x10000 ///< Haswell FMA3 functions 64 #define AV_CPU_FLAG_BMI1 0x20000 ///< Bit Manipulation Instruction Set 1 65 #define AV_CPU_FLAG_BMI2 0x40000 ///< Bit Manipulation Instruction Set 2 66 #define AV_CPU_FLAG_AVX512 \ 67 0x100000 ///< AVX-512 functions: requires OS support even if YMM/ZMM 68 ///< registers aren't used 69 #define AV_CPU_FLAG_SLOW_GATHER 0x2000000 ///< CPU has slow gathers. 70 71 #define AV_CPU_FLAG_ALTIVEC 0x0001 ///< standard 72 #define AV_CPU_FLAG_VSX 0x0002 ///< ISA 2.06 73 #define AV_CPU_FLAG_POWER8 0x0004 ///< ISA 2.07 74 75 #define AV_CPU_FLAG_ARMV5TE (1 << 0) 76 #define AV_CPU_FLAG_ARMV6 (1 << 1) 77 #define AV_CPU_FLAG_ARMV6T2 (1 << 2) 78 #define AV_CPU_FLAG_VFP (1 << 3) 79 #define AV_CPU_FLAG_VFPV3 (1 << 4) 80 #define AV_CPU_FLAG_NEON (1 << 5) 81 #define AV_CPU_FLAG_ARMV8 (1 << 6) 82 #define AV_CPU_FLAG_VFP_VM \ 83 (1 << 7) ///< VFPv2 vector mode, deprecated in ARMv7-A and unavailable in 84 ///< various CPUs implementations 85 #define AV_CPU_FLAG_SETEND (1 << 16) 86 87 #define AV_CPU_FLAG_MMI (1 << 0) 88 #define AV_CPU_FLAG_MSA (1 << 1) 89 90 // Loongarch SIMD extension. 91 #define AV_CPU_FLAG_LSX (1 << 0) 92 #define AV_CPU_FLAG_LASX (1 << 1) 93 94 /** 95 * Return the flags which specify extensions supported by the CPU. 96 * The returned value is affected by av_force_cpu_flags() if that was used 97 * before. So av_get_cpu_flags() can easily be used in an application to 98 * detect the enabled cpu flags. 99 */ 100 int av_get_cpu_flags(void); 101 102 /** 103 * Disables cpu detection and forces the specified flags. 104 * -1 is a special case that disables forcing of specific flags. 105 */ 106 void av_force_cpu_flags(int flags); 107 108 /** 109 * Parse CPU caps from a string and update the given AV_CPU_* flags based on 110 * that. 111 * 112 * @return negative on error. 113 */ 114 int av_parse_cpu_caps(unsigned* flags, const char* s); 115 116 /** 117 * @return the number of logical CPU cores present. 118 */ 119 int av_cpu_count(void); 120 121 /** 122 * Overrides cpu count detection and forces the specified count. 123 * Count < 1 disables forcing of specific count. 124 */ 125 void av_cpu_force_count(int count); 126 127 /** 128 * Get the maximum data alignment that may be required by FFmpeg. 129 * 130 * Note that this is affected by the build configuration and the CPU flags mask, 131 * so e.g. if the CPU supports AVX, but libavutil has been built with 132 * --disable-avx or the AV_CPU_FLAG_AVX flag has been disabled through 133 * av_set_cpu_flags_mask(), then this function will behave as if AVX is not 134 * present. 135 */ 136 size_t av_cpu_max_align(void); 137 138 #endif /* AVUTIL_CPU_H */