ffmpeg/libavcodec/x86/blockdsp.asm

   1 ;******************************************************************************
   2 ;* SIMD-optimized clear block functions
   3 ;* Copyright (c) 2002 Michael Niedermayer
   4 ;* Copyright (c) 2008 Loren Merritt
   5 ;* Copyright (c) 2009 Fiona Glaser
   6 ;*
   7 ;* This file is part of FFmpeg.
   8 ;*
   9 ;* FFmpeg is free software; you can redistribute it and/or
  10 ;* modify it under the terms of the GNU Lesser General Public
  11 ;* License as published by the Free Software Foundation; either
  12 ;* version 2.1 of the License, or (at your option) any later version.
  13 ;*
  14 ;* FFmpeg is distributed in the hope that it will be useful,
  15 ;* but WITHOUT ANY WARRANTY; without even the implied warranty of
  16 ;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17 ;* Lesser General Public License for more details.
  18 ;*
  19 ;* You should have received a copy of the GNU Lesser General Public
  20 ;* License along with FFmpeg; if not, write to the Free Software
  21 ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  22 ;******************************************************************************
  23
  24 %include "libavutil/x86/x86util.asm"
  25
  26 SECTION_TEXT
  27
  28 ;----------------------------------------
  29 ; void ff_clear_block(int16_t *blocks);
  30 ;----------------------------------------
  31 ; %1 = number of xmm registers used
  32 ; %2 = number of inline store loops
  33 %macro CLEAR_BLOCK 2
  34 cglobal clear_block, 1, 1, %1, blocks
  35     ZERO  m0, m0
  36 %assign %%i 0
  37 %rep %2
  38     mova  [blocksq+mmsize*(0+%%i)], m0
  39     mova  [blocksq+mmsize*(1+%%i)], m0
  40     mova  [blocksq+mmsize*(2+%%i)], m0
  41     mova  [blocksq+mmsize*(3+%%i)], m0
  42     mova  [blocksq+mmsize*(4+%%i)], m0
  43     mova  [blocksq+mmsize*(5+%%i)], m0
  44     mova  [blocksq+mmsize*(6+%%i)], m0
  45     mova  [blocksq+mmsize*(7+%%i)], m0
  46 %assign %%i %%i+8
  47 %endrep
  48     RET
  49 %endmacro
  50
  51 INIT_MMX mmx
  52 %define ZERO pxor
  53 CLEAR_BLOCK 0, 2
  54 INIT_XMM sse
  55 %define ZERO xorps
  56 CLEAR_BLOCK 1, 1
  57
  58 ;-----------------------------------------
  59 ; void ff_clear_blocks(int16_t *blocks);
  60 ;-----------------------------------------
  61 ; %1 = number of xmm registers used
  62 %macro CLEAR_BLOCKS 1
  63 cglobal clear_blocks, 1, 2, %1, blocks, len
  64     add   blocksq, 768
  65     mov      lenq, -768
  66     ZERO       m0, m0
  67 .loop
  68     mova  [blocksq+lenq+mmsize*0], m0
  69     mova  [blocksq+lenq+mmsize*1], m0
  70     mova  [blocksq+lenq+mmsize*2], m0
  71     mova  [blocksq+lenq+mmsize*3], m0
  72     mova  [blocksq+lenq+mmsize*4], m0
  73     mova  [blocksq+lenq+mmsize*5], m0
  74     mova  [blocksq+lenq+mmsize*6], m0
  75     mova  [blocksq+lenq+mmsize*7], m0
  76     add   lenq, mmsize*8
  77     js .loop
  78     RET
  79 %endmacro
  80
  81 INIT_MMX mmx
  82 %define ZERO pxor
  83 CLEAR_BLOCKS 0
  84 INIT_XMM sse
  85 %define ZERO xorps
  86 CLEAR_BLOCKS 1