libavcodec/ppc/int_altivec.c

   1 /*
   2  * Copyright (c) 2007 Luca Barbato <lu_zero@gentoo.org>
   3  *
   4  * This file is part of FFmpeg.
   5  *
   6  * FFmpeg is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU Lesser General Public
   8  * License as published by the Free Software Foundation; either
   9  * version 2.1 of the License, or (at your option) any later version.
  10  *
  11  * FFmpeg is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  * Lesser General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU Lesser General Public
  17  * License along with FFmpeg; if not, write to the Free Software
  18  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  19  */
  20
  21 /**
  22  ** @file int_altivec.c
  23  ** integer misc ops.
  24  **/
  25
  26 #include "libavcodec/dsputil.h"
  27
  28 #include "gcc_fixes.h"
  29
  30 #include "dsputil_altivec.h"
  31
  32 static int ssd_int8_vs_int16_altivec(const int8_t *pix1, const int16_t *pix2,
  33                                      int size) {
  34     int i, size16;
  35     vector signed char vpix1;
  36     vector signed short vpix2, vdiff, vpix1l,vpix1h;
  37     union { vector signed int vscore;
  38             int32_t score[4];
  39            } u;
  40     u.vscore = vec_splat_s32(0);
  41 //
  42 //XXX lazy way, fix it later
  43
  44 #define vec_unaligned_load(b) \
  45     vec_perm(vec_ld(0,b),vec_ld(15,b),vec_lvsl(0, b));
  46
  47     size16 = size >> 4;
  48     while(size16) {
  49 //        score += (pix1[i]-pix2[i])*(pix1[i]-pix2[i]);
  50         //load pix1 and the first batch of pix2
  51
  52         vpix1 = vec_unaligned_load(pix1);
  53         vpix2 = vec_unaligned_load(pix2);
  54         pix2 += 8;
  55         //unpack
  56         vpix1h = vec_unpackh(vpix1);
  57         vdiff  = vec_sub(vpix1h, vpix2);
  58         vpix1l = vec_unpackl(vpix1);
  59         // load another batch from pix2
  60         vpix2 = vec_unaligned_load(pix2);
  61         u.vscore = vec_msum(vdiff, vdiff, u.vscore);
  62         vdiff  = vec_sub(vpix1l, vpix2);
  63         u.vscore = vec_msum(vdiff, vdiff, u.vscore);
  64         pix1 += 16;
  65         pix2 += 8;
  66         size16--;
  67     }
  68     u.vscore = vec_sums(u.vscore, vec_splat_s32(0));
  69
  70     size %= 16;
  71     for (i = 0; i < size; i++) {
  72         u.score[3] += (pix1[i]-pix2[i])*(pix1[i]-pix2[i]);
  73     }
  74     return u.score[3];
  75 }
  76
  77 void int_init_altivec(DSPContext* c, AVCodecContext *avctx)
  78 {
  79     c->ssd_int8_vs_int16 = ssd_int8_vs_int16_altivec;
  80 }