2 * Copyright (c) 2015 James Almer
4 * This file is part of FFmpeg.
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.
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.
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
21 #ifndef AVUTIL_X86_INTMATH_H
22 #define AVUTIL_X86_INTMATH_H
29 #elif defined(__INTEL_COMPILER)
30 #include <immintrin.h>
36 #if (defined(__INTEL_COMPILER) && (__INTEL_COMPILER>=1216)) || defined(_MSC_VER)
37 # if defined(__INTEL_COMPILER)
38 # define ff_log2(x) (_bit_scan_reverse((x)|1))
40 # define ff_log2 ff_log2_x86
41 static av_always_inline av_const int ff_log2_x86(unsigned int v)
44 _BitScanReverse(&n, v|1);
48 # define ff_log2_16bit av_log2
50 #if defined(__INTEL_COMPILER) || (defined(_MSC_VER) && (_MSC_VER >= 1700))
51 # define ff_ctz(v) _tzcnt_u32(v)
54 # define ff_ctzll(v) _tzcnt_u64(v)
56 # define ff_ctzll ff_ctzll_x86
57 static av_always_inline av_const int ff_ctzll_x86(long long v)
59 return ((uint32_t)v == 0) ? _tzcnt_u32((uint32_t)(v >> 32)) + 32 : _tzcnt_u32((uint32_t)v);
64 #endif /* __INTEL_COMPILER */
66 #endif /* HAVE_FAST_CLZ */
70 /* Our generic version of av_popcount is faster than GCC's built-in on
71 * CPUs that don't support the popcnt instruction.
73 #if defined(__POPCNT__)
74 #define av_popcount __builtin_popcount
76 #define av_popcount64 __builtin_popcountll
79 #endif /* __POPCNT__ */
83 #if AV_GCC_VERSION_AT_LEAST(5,1)
84 #define av_mod_uintp2 __builtin_ia32_bzhi_si
86 /* GCC releases before 5.1.0 have a broken bzhi builtin, so for those we
87 * implement it using inline assembly
89 #define av_mod_uintp2 av_mod_uintp2_bmi2
90 static av_always_inline av_const unsigned av_mod_uintp2_bmi2(unsigned a, unsigned p)
92 if (av_builtin_constant_p(p))
93 return a & ((1 << p) - 1);
96 __asm__ ("bzhi %2, %1, %0 \n\t" : "=r"(x) : "rm"(a), "r"(p));
100 #endif /* AV_GCC_VERSION_AT_LEAST */
102 #endif /* __BMI2__ */
104 #if defined(__SSE2__) && !defined(__INTEL_COMPILER)
106 #define av_clipd av_clipd_sse2
107 static av_always_inline av_const double av_clipd_sse2(double a, double amin, double amax)
109 #if defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2
110 if (amin > amax) abort();
112 __asm__ ("minsd %2, %0 \n\t"
114 : "+&x"(a) : "xm"(amin), "xm"(amax));
118 #endif /* __SSE2__ */
120 #if defined(__SSE__) && !defined(__INTEL_COMPILER)
122 #define av_clipf av_clipf_sse
123 static av_always_inline av_const float av_clipf_sse(float a, float amin, float amax)
125 #if defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2
126 if (amin > amax) abort();
128 __asm__ ("minss %2, %0 \n\t"
130 : "+&x"(a) : "xm"(amin), "xm"(amax));
136 #endif /* __GNUC__ */
138 #endif /* AVUTIL_X86_INTMATH_H */