Home | History | Annotate | Download | only in x86
      1 /* Copyright (c) 2013 Jean-Marc Valin and John Ridges
      2    Copyright (c) 2014, Cisco Systems, INC MingXiang WeiZhou MinPeng YanWang*/
      3 /**
      4    @file pitch_sse.h
      5    @brief Pitch analysis
      6  */
      7 
      8 /*
      9    Redistribution and use in source and binary forms, with or without
     10    modification, are permitted provided that the following conditions
     11    are met:
     12 
     13    - Redistributions of source code must retain the above copyright
     14    notice, this list of conditions and the following disclaimer.
     15 
     16    - Redistributions in binary form must reproduce the above copyright
     17    notice, this list of conditions and the following disclaimer in the
     18    documentation and/or other materials provided with the distribution.
     19 
     20    THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
     21    ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
     22    LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
     23    A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
     24    OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
     25    EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
     26    PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
     27    PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
     28    LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
     29    NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
     30    SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     31 */
     32 
     33 #ifndef PITCH_SSE_H
     34 #define PITCH_SSE_H
     35 
     36 #if defined(HAVE_CONFIG_H)
     37 #include "config.h"
     38 #endif
     39 
     40 #if defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)
     41 void xcorr_kernel_sse4_1(
     42                     const opus_int16 *x,
     43                     const opus_int16 *y,
     44                     opus_val32       sum[4],
     45                     int              len);
     46 #endif
     47 
     48 #if defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT)
     49 void xcorr_kernel_sse(
     50                     const opus_val16 *x,
     51                     const opus_val16 *y,
     52                     opus_val32       sum[4],
     53                     int              len);
     54 #endif
     55 
     56 #if defined(OPUS_X86_PRESUME_SSE4_1) && defined(FIXED_POINT)
     57 #define OVERRIDE_XCORR_KERNEL
     58 #define xcorr_kernel(x, y, sum, len, arch) \
     59     ((void)arch, xcorr_kernel_sse4_1(x, y, sum, len))
     60 
     61 #elif defined(OPUS_X86_PRESUME_SSE) && !defined(FIXED_POINT)
     62 #define OVERRIDE_XCORR_KERNEL
     63 #define xcorr_kernel(x, y, sum, len, arch) \
     64     ((void)arch, xcorr_kernel_sse(x, y, sum, len))
     65 
     66 #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                     const opus_val16 *x,
     70                     const opus_val16 *y,
     71                     opus_val32       sum[4],
     72                     int              len);
     73 
     74 #define OVERRIDE_XCORR_KERNEL
     75 #define xcorr_kernel(x, y, sum, len, arch) \
     76     ((*XCORR_KERNEL_IMPL[(arch) & OPUS_ARCHMASK])(x, y, sum, len))
     77 
     78 #endif
     79 
     80 #if defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)
     81 opus_val32 celt_inner_prod_sse4_1(
     82     const opus_int16 *x,
     83     const opus_int16 *y,
     84     int               N);
     85 #endif
     86 
     87 #if defined(OPUS_X86_MAY_HAVE_SSE2) && defined(FIXED_POINT)
     88 opus_val32 celt_inner_prod_sse2(
     89     const opus_int16 *x,
     90     const opus_int16 *y,
     91     int               N);
     92 #endif
     93 
     94 #if defined(OPUS_X86_MAY_HAVE_SSE2) && !defined(FIXED_POINT)
     95 opus_val32 celt_inner_prod_sse(
     96     const opus_val16 *x,
     97     const opus_val16 *y,
     98     int               N);
     99 #endif
    100 
    101 
    102 #if defined(OPUS_X86_PRESUME_SSE4_1) && defined(FIXED_POINT)
    103 #define OVERRIDE_CELT_INNER_PROD
    104 #define celt_inner_prod(x, y, N, arch) \
    105     ((void)arch, celt_inner_prod_sse4_1(x, y, N))
    106 
    107 #elif defined(OPUS_X86_PRESUME_SSE2) && defined(FIXED_POINT) && !defined(OPUS_X86_MAY_HAVE_SSE4_1)
    108 #define OVERRIDE_CELT_INNER_PROD
    109 #define celt_inner_prod(x, y, N, arch) \
    110     ((void)arch, celt_inner_prod_sse2(x, y, N))
    111 
    112 #elif defined(OPUS_X86_PRESUME_SSE) && !defined(FIXED_POINT)
    113 #define OVERRIDE_CELT_INNER_PROD
    114 #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                     const opus_val16 *x,
    123                     const opus_val16 *y,
    124                     int               N);
    125 
    126 #define OVERRIDE_CELT_INNER_PROD
    127 #define celt_inner_prod(x, y, N, arch) \
    128     ((*CELT_INNER_PROD_IMPL[(arch) & OPUS_ARCHMASK])(x, y, N))
    129 
    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 #undef dual_inner_prod
    138 #undef comb_filter_const
    139 
    140 void dual_inner_prod_sse(const opus_val16 *x,
    141     const opus_val16 *y01,
    142     const opus_val16 *y02,
    143     int               N,
    144     opus_val32       *xy1,
    145     opus_val32       *xy2);
    146 
    147 void comb_filter_const_sse(opus_val32 *y,
    148     opus_val32 *x,
    149     int         T,
    150     int         N,
    151     opus_val16  g10,
    152     opus_val16  g11,
    153     opus_val16  g12);
    154 
    155 
    156 #if defined(OPUS_X86_PRESUME_SSE)
    157 # define dual_inner_prod(x, y01, y02, N, xy1, xy2, arch) \
    158     ((void)(arch),dual_inner_prod_sse(x, y01, y02, N, xy1, xy2))
    159 
    160 # define comb_filter_const(y, x, T, N, g10, g11, g12, arch) \
    161     ((void)(arch),comb_filter_const_sse(y, x, T, N, g10, g11, g12))
    162 #else
    163 
    164 extern void (*const DUAL_INNER_PROD_IMPL[OPUS_ARCHMASK + 1])(
    165               const opus_val16 *x,
    166               const opus_val16 *y01,
    167               const opus_val16 *y02,
    168               int               N,
    169               opus_val32       *xy1,
    170               opus_val32       *xy2);
    171 
    172 #define dual_inner_prod(x, y01, y02, N, xy1, xy2, arch) \
    173     ((*DUAL_INNER_PROD_IMPL[(arch) & OPUS_ARCHMASK])(x, y01, y02, N, xy1, xy2))
    174 
    175 extern void (*const COMB_FILTER_CONST_IMPL[OPUS_ARCHMASK + 1])(
    176               opus_val32 *y,
    177               opus_val32 *x,
    178               int         T,
    179               int         N,
    180               opus_val16  g10,
    181               opus_val16  g11,
    182               opus_val16  g12);
    183 
    184 #define comb_filter_const(y, x, T, N, g10, g11, g12, arch) \
    185     ((*COMB_FILTER_CONST_IMPL[(arch) & OPUS_ARCHMASK])(y, x, T, N, g10, g11, g12))
    186 
    187 #define NON_STATIC_COMB_FILTER_CONST_C
    188 
    189 #endif
    190 #endif
    191 
    192 #endif
    193