libavcodec/x86/fpel.asm

   1 ;******************************************************************************
   2 ;* SIMD-optimized fullpel functions
   3 ;* Copyright (c) 2008 Loren Merritt
   4 ;* Copyright (c) 2003-2013 Michael Niedermayer
   5 ;* Copyright (c) 2013 Daniel Kang
   6 ;*
   7 ;* This file is part of Libav.
   8 ;*
   9 ;* Libav 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 ;* Libav 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 Libav; 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 INIT_MMX mmxext
  29 ; void ff_put/avg_pixels(uint8_t *block, const uint8_t *pixels,
  30 ;                        ptrdiff_t line_size, int h)
  31 %macro PIXELS48 2
  32 %if %2 == 4
  33 %define OP movh
  34 %else
  35 %define OP mova
  36 %endif
  37 cglobal %1_pixels%2, 4,5
  38     lea          r4, [r2*3]
  39 .loop:
  40     OP           m0, [r1]
  41     OP           m1, [r1+r2]
  42     OP           m2, [r1+r2*2]
  43     OP           m3, [r1+r4]
  44     lea          r1, [r1+r2*4]
  45 %ifidn %1, avg
  46     pavgb        m0, [r0]
  47     pavgb        m1, [r0+r2]
  48     pavgb        m2, [r0+r2*2]
  49     pavgb        m3, [r0+r4]
  50 %endif
  51     OP         [r0], m0
  52     OP      [r0+r2], m1
  53     OP    [r0+r2*2], m2
  54     OP      [r0+r4], m3
  55     sub         r3d, 4
  56     lea          r0, [r0+r2*4]
  57     jne       .loop
  58     RET
  59 %endmacro
  60
  61 PIXELS48 put, 4
  62 PIXELS48 avg, 4
  63 PIXELS48 put, 8
  64 PIXELS48 avg, 8
  65
  66
  67 INIT_XMM sse2
  68 ; void ff_put_pixels16_sse2(uint8_t *block, const uint8_t *pixels,
  69 ;                           ptrdiff_t line_size, int h)
  70 cglobal put_pixels16, 4,5,4
  71     lea          r4, [r2*3]
  72 .loop:
  73     movu         m0, [r1]
  74     movu         m1, [r1+r2]
  75     movu         m2, [r1+r2*2]
  76     movu         m3, [r1+r4]
  77     lea          r1, [r1+r2*4]
  78     mova       [r0], m0
  79     mova    [r0+r2], m1
  80     mova  [r0+r2*2], m2
  81     mova    [r0+r4], m3
  82     sub         r3d, 4
  83     lea          r0, [r0+r2*4]
  84     jnz       .loop
  85     REP_RET
  86
  87 ; void ff_avg_pixels16_sse2(uint8_t *block, const uint8_t *pixels,
  88 ;                           ptrdiff_t line_size, int h)
  89 cglobal avg_pixels16, 4,5,4
  90     lea          r4, [r2*3]
  91 .loop:
  92     movu         m0, [r1]
  93     movu         m1, [r1+r2]
  94     movu         m2, [r1+r2*2]
  95     movu         m3, [r1+r4]
  96     lea          r1, [r1+r2*4]
  97     pavgb        m0, [r0]
  98     pavgb        m1, [r0+r2]
  99     pavgb        m2, [r0+r2*2]
 100     pavgb        m3, [r0+r4]
 101     mova       [r0], m0
 102     mova    [r0+r2], m1
 103     mova  [r0+r2*2], m2
 104     mova    [r0+r4], m3
 105     sub         r3d, 4
 106     lea          r0, [r0+r2*4]
 107     jnz       .loop
 108     REP_RET