libavcodec/x86/dnxhd_mmx.c

   1 /*
   2  * VC3/DNxHD SIMD functions
   3  * Copyright (c) 2007 Baptiste Coudurier <baptiste dot coudurier at smartjog dot com>
   4  *
   5  * VC-3 encoder funded by the British Broadcasting Corporation
   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_cpu.h"
  25 #include "libavcodec/dnxhdenc.h"
  26
  27 static void get_pixels_8x4_sym_sse2(DCTELEM *block, const uint8_t *pixels, int line_size)
  28 {
  29     __asm__ volatile(
  30         "pxor %%xmm7,      %%xmm7       \n\t"
  31         "movq (%0),        %%xmm0       \n\t"
  32         "add  %2,          %0           \n\t"
  33         "movq (%0),        %%xmm1       \n\t"
  34         "movq (%0, %2),    %%xmm2       \n\t"
  35         "movq (%0, %2,2),  %%xmm3       \n\t"
  36         "punpcklbw %%xmm7, %%xmm0       \n\t"
  37         "punpcklbw %%xmm7, %%xmm1       \n\t"
  38         "punpcklbw %%xmm7, %%xmm2       \n\t"
  39         "punpcklbw %%xmm7, %%xmm3       \n\t"
  40         "movdqa %%xmm0,      (%1)       \n\t"
  41         "movdqa %%xmm1,    16(%1)       \n\t"
  42         "movdqa %%xmm2,    32(%1)       \n\t"
  43         "movdqa %%xmm3,    48(%1)       \n\t"
  44         "movdqa %%xmm3 ,   64(%1)       \n\t"
  45         "movdqa %%xmm2 ,   80(%1)       \n\t"
  46         "movdqa %%xmm1 ,   96(%1)       \n\t"
  47         "movdqa %%xmm0,   112(%1)       \n\t"
  48         : "+r" (pixels)
  49         : "r" (block), "r" ((x86_reg)line_size)
  50     );
  51 }
  52
  53 void ff_dnxhd_init_mmx(DNXHDEncContext *ctx)
  54 {
  55     if (mm_flags & FF_MM_SSE2) {
  56         ctx->get_pixels_8x4_sym = get_pixels_8x4_sym_sse2;
  57     }
  58 }