libavfilter/x86/vf_overlay.asm

   1 ;*****************************************************************************
   2 ;* x86-optimized functions for overlay filter
   3 ;*
   4 ;* Copyright (C) 2018 Paul B Mahol
   5 ;* Copyright (C) 2018 Henrik Gramner
   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_RODATA
  27
  28 pb_1:     times 16 db 1
  29 pw_128:   times  8 dw 128
  30 pw_255:   times  8 dw 255
  31 pw_257:   times  8 dw 257
  32
  33 SECTION .text
  34
  35 INIT_XMM sse4
  36 cglobal overlay_row_44, 5, 7, 6, 0, dst, da, s, a, w, r, x
  37     xor          xq, xq
  38     movsxdifnidn wq, wd
  39     mov          rq, wq
  40     and          rq, mmsize/2 - 1
  41     cmp          wq, mmsize/2
  42     jl .end
  43     sub          wq, rq
  44     mova         m3, [pw_255]
  45     mova         m4, [pw_128]
  46     mova         m5, [pw_257]
  47     .loop:
  48         pmovzxbw    m0, [sq+xq]
  49         pmovzxbw    m2, [aq+xq]
  50         pmovzxbw    m1, [dstq+xq]
  51         pmullw      m0, m2
  52         pxor        m2, m3
  53         pmullw      m1, m2
  54         paddw       m0, m4
  55         paddw       m0, m1
  56         pmulhuw     m0, m5
  57         packuswb    m0, m0
  58         movq [dstq+xq], m0
  59         add         xq, mmsize/2
  60         cmp         xq, wq
  61         jl .loop
  62
  63     .end:
  64     mov    eax, xd
  65     RET
  66
  67 INIT_XMM sse4
  68 cglobal overlay_row_22, 5, 7, 6, 0, dst, da, s, a, w, r, x
  69     xor          xq, xq
  70     movsxdifnidn wq, wd
  71     sub          wq, 1
  72     mov          rq, wq
  73     and          rq, mmsize/2 - 1
  74     cmp          wq, mmsize/2
  75     jl .end
  76     sub          wq, rq
  77     mova         m3, [pw_255]
  78     mova         m4, [pw_128]
  79     mova         m5, [pw_257]
  80     .loop:
  81         pmovzxbw    m0, [sq+xq]
  82         movu        m1, [aq+2*xq]
  83         pandn       m2, m3, m1
  84         psllw       m1, 8
  85         pavgw       m2, m1
  86         pavgw       m2, m1
  87         psrlw       m2, 8
  88         pmovzxbw    m1, [dstq+xq]
  89         pmullw      m0, m2
  90         pxor        m2, m3
  91         pmullw      m1, m2
  92         paddw       m0, m4
  93         paddw       m0, m1
  94         pmulhuw     m0, m5
  95         packuswb    m0, m0
  96         movq [dstq+xq], m0
  97         add         xq, mmsize/2
  98         cmp         xq, wq
  99         jl .loop
 100
 101     .end:
 102     mov    eax, xd
 103     RET
 104
 105 INIT_XMM sse4
 106 cglobal overlay_row_20, 6, 7, 7, 0, dst, da, s, a, w, r, x
 107     mov         daq, aq
 108     add         daq, rmp
 109     xor          xq, xq
 110     movsxdifnidn wq, wd
 111     sub          wq, 1
 112     mov          rq, wq
 113     and          rq, mmsize/2 - 1
 114     cmp          wq, mmsize/2
 115     jl .end
 116     sub          wq, rq
 117     mova         m3, [pw_255]
 118     mova         m4, [pw_128]
 119     mova         m5, [pw_257]
 120     mova         m6, [pb_1]
 121     .loop:
 122         pmovzxbw    m0, [sq+xq]
 123         movu        m2, [aq+2*xq]
 124         movu        m1, [daq+2*xq]
 125         pmaddubsw   m2, m6
 126         pmaddubsw   m1, m6
 127         paddw       m2, m1
 128         psrlw       m2, 2
 129         pmovzxbw    m1, [dstq+xq]
 130         pmullw      m0, m2
 131         pxor        m2, m3
 132         pmullw      m1, m2
 133         paddw       m0, m4
 134         paddw       m0, m1
 135         pmulhuw     m0, m5
 136         packuswb    m0, m0
 137         movq [dstq+xq], m0
 138         add         xq, mmsize/2
 139         cmp         xq, wq
 140         jl .loop
 141
 142     .end:
 143     mov    eax, xd
 144     RET