cpu.h (6769B)
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 #include "version.h" 26 27 #define AV_CPU_FLAG_FORCE 0x80000000 /* force usage of selected flags (OR) */ 28 29 /* lower 16 bits - CPU features */ 30 #define AV_CPU_FLAG_MMX 0x0001 ///< standard MMX 31 #define AV_CPU_FLAG_MMXEXT 0x0002 ///< SSE integer functions or AMD MMX ext 32 #define AV_CPU_FLAG_MMX2 0x0002 ///< SSE integer functions or AMD MMX ext 33 #define AV_CPU_FLAG_3DNOW 0x0004 ///< AMD 3DNOW 34 #define AV_CPU_FLAG_SSE 0x0008 ///< SSE functions 35 #define AV_CPU_FLAG_SSE2 0x0010 ///< PIV SSE2 functions 36 #define AV_CPU_FLAG_SSE2SLOW 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 0x20000000 ///< SSE3 supported, but usually not faster 41 ///< than regular MMX/SSE (e.g. Core1) 42 #define AV_CPU_FLAG_SSSE3 0x0080 ///< Conroe SSSE3 functions 43 #define AV_CPU_FLAG_SSSE3SLOW 0x4000000 ///< SSSE3 supported, but usually not faster 44 #define AV_CPU_FLAG_ATOM 0x10000000 ///< Atom processor, some SSSE3 instructions are slower 45 #define AV_CPU_FLAG_SSE4 0x0100 ///< Penryn SSE4.1 functions 46 #define AV_CPU_FLAG_SSE42 0x0200 ///< Nehalem SSE4.2 functions 47 #define AV_CPU_FLAG_AESNI 0x80000 ///< Advanced Encryption Standard functions 48 #define AV_CPU_FLAG_AVX 0x4000 ///< AVX functions: requires OS support even if YMM registers aren't used 49 #define AV_CPU_FLAG_AVXSLOW 0x8000000 ///< AVX supported, but slow when using YMM registers (e.g. Bulldozer) 50 #define AV_CPU_FLAG_XOP 0x0400 ///< Bulldozer XOP functions 51 #define AV_CPU_FLAG_FMA4 0x0800 ///< Bulldozer FMA4 functions 52 #define AV_CPU_FLAG_CMOV 0x1000 ///< supports cmov instruction 53 #define AV_CPU_FLAG_AVX2 0x8000 ///< AVX2 functions: requires OS support even if YMM registers aren't used 54 #define AV_CPU_FLAG_FMA3 0x10000 ///< Haswell FMA3 functions 55 #define AV_CPU_FLAG_BMI1 0x20000 ///< Bit Manipulation Instruction Set 1 56 #define AV_CPU_FLAG_BMI2 0x40000 ///< Bit Manipulation Instruction Set 2 57 #define AV_CPU_FLAG_AVX512 0x100000 ///< AVX-512 functions: requires OS support even if YMM/ZMM registers aren't used 58 #define AV_CPU_FLAG_AVX512ICL 0x200000 ///< F/CD/BW/DQ/VL/VNNI/IFMA/VBMI/VBMI2/VPOPCNTDQ/BITALG/GFNI/VAES/VPCLMULQDQ 59 #define AV_CPU_FLAG_SLOW_GATHER 0x2000000 ///< CPU has slow gathers. 60 61 #define AV_CPU_FLAG_ALTIVEC 0x0001 ///< standard 62 #define AV_CPU_FLAG_VSX 0x0002 ///< ISA 2.06 63 #define AV_CPU_FLAG_POWER8 0x0004 ///< ISA 2.07 64 65 #define AV_CPU_FLAG_ARMV5TE (1 << 0) 66 #define AV_CPU_FLAG_ARMV6 (1 << 1) 67 #define AV_CPU_FLAG_ARMV6T2 (1 << 2) 68 #define AV_CPU_FLAG_VFP (1 << 3) 69 #define AV_CPU_FLAG_VFPV3 (1 << 4) 70 #define AV_CPU_FLAG_NEON (1 << 5) 71 #define AV_CPU_FLAG_ARMV8 (1 << 6) 72 #define AV_CPU_FLAG_VFP_VM (1 << 7) ///< VFPv2 vector mode, deprecated in ARMv7-A and unavailable in various CPUs implementations 73 #define AV_CPU_FLAG_DOTPROD (1 << 8) 74 #define AV_CPU_FLAG_I8MM (1 << 9) 75 #define AV_CPU_FLAG_SVE (1 <<10) 76 #define AV_CPU_FLAG_SVE2 (1 <<11) 77 #define AV_CPU_FLAG_SETEND (1 <<16) 78 79 #define AV_CPU_FLAG_MMI (1 << 0) 80 #define AV_CPU_FLAG_MSA (1 << 1) 81 82 //Loongarch SIMD extension. 83 #define AV_CPU_FLAG_LSX (1 << 0) 84 #define AV_CPU_FLAG_LASX (1 << 1) 85 86 // RISC-V extensions 87 #define AV_CPU_FLAG_RVI (1 << 0) ///< I (full GPR bank) 88 #if FF_API_RISCV_FD_ZBA 89 #define AV_CPU_FLAG_RVF (1 << 1) ///< F (single precision FP) 90 #define AV_CPU_FLAG_RVD (1 << 2) ///< D (double precision FP) 91 #endif 92 #define AV_CPU_FLAG_RVV_I32 (1 << 3) ///< Vectors of 8/16/32-bit int's */ 93 #define AV_CPU_FLAG_RVV_F32 (1 << 4) ///< Vectors of float's */ 94 #define AV_CPU_FLAG_RVV_I64 (1 << 5) ///< Vectors of 64-bit int's */ 95 #define AV_CPU_FLAG_RVV_F64 (1 << 6) ///< Vectors of double's 96 #define AV_CPU_FLAG_RVB_BASIC (1 << 7) ///< Basic bit-manipulations 97 #if FF_API_RISCV_FD_ZBA 98 #define AV_CPU_FLAG_RVB_ADDR (1 << 8) ///< Address bit-manipulations 99 #endif 100 #define AV_CPU_FLAG_RV_ZVBB (1 << 9) ///< Vector basic bit-manipulations 101 #define AV_CPU_FLAG_RV_MISALIGNED (1 <<10) ///< Fast misaligned accesses 102 #define AV_CPU_FLAG_RVB (1 <<11) ///< B (bit manipulations) 103 104 // WASM extensions 105 #define AV_CPU_FLAG_SIMD128 (1 << 0) 106 107 /** 108 * Return the flags which specify extensions supported by the CPU. 109 * The returned value is affected by av_force_cpu_flags() if that was used 110 * before. So av_get_cpu_flags() can easily be used in an application to 111 * detect the enabled cpu flags. 112 */ 113 int av_get_cpu_flags(void); 114 115 /** 116 * Disables cpu detection and forces the specified flags. 117 * -1 is a special case that disables forcing of specific flags. 118 */ 119 void av_force_cpu_flags(int flags); 120 121 /** 122 * Parse CPU caps from a string and update the given AV_CPU_* flags based on that. 123 * 124 * @return negative on error. 125 */ 126 int av_parse_cpu_caps(unsigned *flags, const char *s); 127 128 /** 129 * @return the number of logical CPU cores present. 130 */ 131 int av_cpu_count(void); 132 133 /** 134 * Overrides cpu count detection and forces the specified count. 135 * Count < 1 disables forcing of specific count. 136 */ 137 void av_cpu_force_count(int count); 138 139 /** 140 * Get the maximum data alignment that may be required by FFmpeg. 141 * 142 * Note that this is affected by the build configuration and the CPU flags mask, 143 * so e.g. if the CPU supports AVX, but libavutil has been built with 144 * --disable-avx or the AV_CPU_FLAG_AVX flag has been disabled through 145 * av_set_cpu_flags_mask(), then this function will behave as if AVX is not 146 * present. 147 */ 148 size_t av_cpu_max_align(void); 149 150 #endif /* AVUTIL_CPU_H */