pitch_sse.h 6.21 KB
Newer Older
1 2
/* Copyright (c) 2013 Jean-Marc Valin and John Ridges
   Copyright (c) 2014, Cisco Systems, INC MingXiang WeiZhou MinPeng YanWang*/
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
/**
   @file pitch_sse.h
   @brief Pitch analysis
 */

/*
   Redistribution and use in source and binary forms, with or without
   modification, are permitted provided that the following conditions
   are met:

   - Redistributions of source code must retain the above copyright
   notice, this list of conditions and the following disclaimer.

   - Redistributions in binary form must reproduce the above copyright
   notice, this list of conditions and the following disclaimer in the
   documentation and/or other materials provided with the distribution.

   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
   ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
   OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
   EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
   PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
   PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
   LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
   NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
   SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/

#ifndef PITCH_SSE_H
#define PITCH_SSE_H

36 37 38 39
#if defined(HAVE_CONFIG_H)
#include "config.h"
#endif

40
#if defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)
41 42 43 44
void xcorr_kernel_sse4_1(
                    const opus_int16 *x,
                    const opus_int16 *y,
                    opus_val32       sum[4],
45
                    int              len);
46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
#endif

#if defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT)
void xcorr_kernel_sse(
                    const opus_val16 *x,
                    const opus_val16 *y,
                    opus_val32       sum[4],
                    int              len);
#endif

#if defined(OPUS_X86_PRESUME_SSE4_1) && defined(FIXED_POINT)
#define OVERRIDE_XCORR_KERNEL
#define xcorr_kernel(x, y, sum, len, arch) \
    ((void)arch, xcorr_kernel_sse4_1(x, y, sum, len))

#elif defined(OPUS_X86_PRESUME_SSE) && !defined(FIXED_POINT)
#define OVERRIDE_XCORR_KERNEL
#define xcorr_kernel(x, y, sum, len, arch) \
    ((void)arch, xcorr_kernel_sse(x, y, sum, len))

#elif (defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)) || (defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT))
67 68

extern void (*const XCORR_KERNEL_IMPL[OPUS_ARCHMASK + 1])(
69 70
                    const opus_val16 *x,
                    const opus_val16 *y,
71
                    opus_val32       sum[4],
72
                    int              len);
73

74
#define OVERRIDE_XCORR_KERNEL
75 76 77
#define xcorr_kernel(x, y, sum, len, arch) \
    ((*XCORR_KERNEL_IMPL[(arch) & OPUS_ARCHMASK])(x, y, sum, len))

78 79 80
#endif

#if defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)
81 82 83 84 85 86
opus_val32 celt_inner_prod_sse4_1(
    const opus_int16 *x,
    const opus_int16 *y,
    int               N);
#endif

87
#if defined(OPUS_X86_MAY_HAVE_SSE2) && defined(FIXED_POINT)
88 89 90 91 92 93
opus_val32 celt_inner_prod_sse2(
    const opus_int16 *x,
    const opus_int16 *y,
    int               N);
#endif

94 95 96 97 98 99 100 101 102 103 104
#if defined(OPUS_X86_MAY_HAVE_SSE2) && !defined(FIXED_POINT)
opus_val32 celt_inner_prod_sse(
    const opus_val16 *x,
    const opus_val16 *y,
    int               N);
#endif


#if defined(OPUS_X86_PRESUME_SSE4_1) && defined(FIXED_POINT)
#define OVERRIDE_CELT_INNER_PROD
#define celt_inner_prod(x, y, N, arch) \
105
    ((void)arch, celt_inner_prod_sse4_1(x, y, N))
106 107 108 109

#elif defined(OPUS_X86_PRESUME_SSE2) && defined(FIXED_POINT) && !defined(OPUS_X86_MAY_HAVE_SSE4_1)
#define OVERRIDE_CELT_INNER_PROD
#define celt_inner_prod(x, y, N, arch) \
110
    ((void)arch, celt_inner_prod_sse2(x, y, N))
111 112 113 114

#elif defined(OPUS_X86_PRESUME_SSE) && !defined(FIXED_POINT)
#define OVERRIDE_CELT_INNER_PROD
#define celt_inner_prod(x, y, N, arch) \
115
    ((void)arch, celt_inner_prod_sse(x, y, N))
116 117 118


#elif ((defined(OPUS_X86_MAY_HAVE_SSE4_1) || defined(OPUS_X86_MAY_HAVE_SSE2)) && defined(FIXED_POINT)) || \
119
    (defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT))
120

121
extern opus_val32 (*const CELT_INNER_PROD_IMPL[OPUS_ARCHMASK + 1])(
122 123
                    const opus_val16 *x,
                    const opus_val16 *y,
124 125 126 127 128 129
                    int               N);

#define OVERRIDE_CELT_INNER_PROD
#define celt_inner_prod(x, y, N, arch) \
    ((*CELT_INNER_PROD_IMPL[(arch) & OPUS_ARCHMASK])(x, y, N))

130
#endif
131

132
#if defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT)
133

134
#define OVERRIDE_DUAL_INNER_PROD
135
#define OVERRIDE_COMB_FILTER_CONST
136

137 138
#undef dual_inner_prod
#undef comb_filter_const
139

140
void dual_inner_prod_sse(const opus_val16 *x,
141 142 143 144 145
    const opus_val16 *y01,
    const opus_val16 *y02,
    int               N,
    opus_val32       *xy1,
    opus_val32       *xy2);
146 147

void comb_filter_const_sse(opus_val32 *y,
148 149 150 151 152 153
    opus_val32 *x,
    int         T,
    int         N,
    opus_val16  g10,
    opus_val16  g11,
    opus_val16  g12);
154

155 156 157 158 159 160 161

#if defined(OPUS_X86_PRESUME_SSE)
# define dual_inner_prod(x, y01, y02, N, xy1, xy2, arch) \
    ((void)(arch),dual_inner_prod_sse(x, y01, y02, N, xy1, xy2))

# define comb_filter_const(y, x, T, N, g10, g11, g12, arch) \
    ((void)(arch),comb_filter_const_sse(y, x, T, N, g10, g11, g12))
162
#else
163 164 165 166 167 168 169 170 171

extern void (*const DUAL_INNER_PROD_IMPL[OPUS_ARCHMASK + 1])(
              const opus_val16 *x,
              const opus_val16 *y01,
              const opus_val16 *y02,
              int               N,
              opus_val32       *xy1,
              opus_val32       *xy2);

172
#define dual_inner_prod(x, y01, y02, N, xy1, xy2, arch) \
173 174 175 176 177 178 179 180 181 182 183
    ((*DUAL_INNER_PROD_IMPL[(arch) & OPUS_ARCHMASK])(x, y01, y02, N, xy1, xy2))

extern void (*const COMB_FILTER_CONST_IMPL[OPUS_ARCHMASK + 1])(
              opus_val32 *y,
              opus_val32 *x,
              int         T,
              int         N,
              opus_val16  g10,
              opus_val16  g11,
              opus_val16  g12);

184
#define comb_filter_const(y, x, T, N, g10, g11, g12, arch) \
185 186 187 188
    ((*COMB_FILTER_CONST_IMPL[(arch) & OPUS_ARCHMASK])(y, x, T, N, g10, g11, g12))

#define NON_STATIC_COMB_FILTER_CONST_C

189
#endif
190
#endif
191

192
#endif