ffmpeg/libavcodec/x86/hevc_idct.asm

   1 ; /*
   2 ; * SIMD optimized idct functions for HEVC decoding
   3 ; * Copyright (c) 2014 Pierre-Edouard LEPERE
   4 ; * Copyright (c) 2014 James Almer
   5 ; *
   6 ; * This file is part of FFmpeg.
   7 ; *
   8 ; * FFmpeg is free software; you can redistribute it and/or
   9 ; * modify it under the terms of the GNU Lesser General Public
  10 ; * License as published by the Free Software Foundation; either
  11 ; * version 2.1 of the License, or (at your option) any later version.
  12 ; *
  13 ; * FFmpeg is distributed in the hope that it will be useful,
  14 ; * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15 ; * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  16 ; * Lesser General Public License for more details.
  17 ; *
  18 ; * You should have received a copy of the GNU Lesser General Public
  19 ; * License along with FFmpeg; if not, write to the Free Software
  20 ; * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  21 ; */
  22 %include "libavutil/x86/x86util.asm"
  23
  24 SECTION_TEXT 32
  25
  26 ; void ff_hevc_idctHxW_dc_{8,10}_<opt>(int16_t *coeffs)
  27 ; %1 = HxW
  28 ; %2 = number of loops
  29 ; %3 = bitdepth
  30 %macro IDCT_DC 3
  31 cglobal hevc_idct%1x%1_dc_%3, 1, 2, 1, coeff, tmp
  32     movsx             tmpq, word [coeffq]
  33     add               tmpw, ((1 << 14-%3) + 1)
  34     sar               tmpw, (15-%3)
  35     movd               xm0, tmpd
  36     SPLATW              m0, xm0
  37     DEFINE_ARGS coeff, cnt
  38     mov               cntd, %2
  39 .loop:
  40     mova [coeffq+mmsize*0], m0
  41     mova [coeffq+mmsize*1], m0
  42     mova [coeffq+mmsize*2], m0
  43     mova [coeffq+mmsize*3], m0
  44     mova [coeffq+mmsize*4], m0
  45     mova [coeffq+mmsize*5], m0
  46     mova [coeffq+mmsize*6], m0
  47     mova [coeffq+mmsize*7], m0
  48     add  coeffq, mmsize*8
  49     dec  cntd
  50     jg  .loop
  51     RET
  52 %endmacro
  53
  54 ; %1 = HxW
  55 ; %2 = bitdepth
  56 %macro IDCT_DC_NL 2 ; No loop
  57 cglobal hevc_idct%1x%1_dc_%2, 1, 2, 1, coeff, tmp
  58     movsx             tmpq, word [coeffq]
  59     add               tmpw, ((1 << 14-%2) + 1)
  60     sar               tmpw, (15-%2)
  61     movd                m0, tmpd
  62     SPLATW              m0, xm0
  63     mova [coeffq+mmsize*0], m0
  64     mova [coeffq+mmsize*1], m0
  65     mova [coeffq+mmsize*2], m0
  66     mova [coeffq+mmsize*3], m0
  67 %if mmsize == 16
  68     mova [coeffq+mmsize*4], m0
  69     mova [coeffq+mmsize*5], m0
  70     mova [coeffq+mmsize*6], m0
  71     mova [coeffq+mmsize*7], m0
  72 %endif
  73     RET
  74 %endmacro
  75
  76 ; 8-bit
  77 INIT_MMX mmxext
  78 IDCT_DC_NL  4,      8
  79 IDCT_DC     8,  2,  8
  80
  81 INIT_XMM sse2
  82 IDCT_DC_NL  8,      8
  83 IDCT_DC    16,  4,  8
  84 IDCT_DC    32, 16,  8
  85
  86 %if HAVE_AVX2_EXTERNAL
  87 INIT_YMM avx2
  88 IDCT_DC    16,  2,  8
  89 IDCT_DC    32,  8,  8
  90 %endif ;HAVE_AVX2_EXTERNAL
  91
  92 ; 10-bit
  93 INIT_MMX mmxext
  94 IDCT_DC_NL  4,     10
  95 IDCT_DC     8,  2, 10
  96
  97 INIT_XMM sse2
  98 IDCT_DC_NL  8,     10
  99 IDCT_DC    16,  4, 10
 100 IDCT_DC    32, 16, 10
 101
 102 %if HAVE_AVX2_EXTERNAL
 103 INIT_YMM avx2
 104 IDCT_DC    16,  2, 10
 105 IDCT_DC    32,  8, 10
 106 %endif ;HAVE_AVX2_EXTERNAL
 107
 108 ; 12-bit
 109 INIT_MMX mmxext
 110 IDCT_DC_NL  4,     12
 111 IDCT_DC     8,  2, 12
 112
 113 INIT_XMM sse2
 114 IDCT_DC_NL  8,     12
 115 IDCT_DC    16,  4, 12
 116 IDCT_DC    32, 16, 12
 117
 118 %if HAVE_AVX2_EXTERNAL
 119 INIT_YMM avx2
 120 IDCT_DC    16,  2, 12
 121 IDCT_DC    32,  8, 12
 122 %endif ;HAVE_AVX2_EXTERNAL