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 #include "config.h"
22
23 #include <stdint.h>
24
25 #include "libavutil/attributes.h"
26 #include "libavutil/cpu.h"
27 #include "libavutil/ppc/cpu.h"
28 #include "libavutil/ppc/util_altivec.h"
29
30 #include "libavcodec/svq1enc.h"
31
32 #if HAVE_ALTIVEC
ssd_int8_vs_int16_altivec(const int8_t * pix1,const int16_t * pix2,intptr_t size)33 static int ssd_int8_vs_int16_altivec(const int8_t *pix1, const int16_t *pix2,
34 intptr_t size)
35 {
36 int i, size16 = size >> 4;
37 vector signed char vpix1;
38 vector signed short vpix2, vdiff, vpix1l, vpix1h;
39 union {
40 vector signed int vscore;
41 int32_t score[4];
42 } u = { .vscore = vec_splat_s32(0) };
43
44 while (size16) {
45 // score += (pix1[i] - pix2[i]) * (pix1[i] - pix2[i]);
46 // load pix1 and the first batch of pix2
47
48 vpix1 = vec_unaligned_load(pix1);
49 vpix2 = vec_unaligned_load(pix2);
50 pix2 += 8;
51 // unpack
52 vpix1h = vec_unpackh(vpix1);
53 vdiff = vec_sub(vpix1h, vpix2);
54 vpix1l = vec_unpackl(vpix1);
55 // load another batch from pix2
56 vpix2 = vec_unaligned_load(pix2);
57 u.vscore = vec_msum(vdiff, vdiff, u.vscore);
58 vdiff = vec_sub(vpix1l, vpix2);
59 u.vscore = vec_msum(vdiff, vdiff, u.vscore);
60 pix1 += 16;
61 pix2 += 8;
62 size16--;
63 }
64 u.vscore = vec_sums(u.vscore, vec_splat_s32(0));
65
66 size %= 16;
67 for (i = 0; i < size; i++)
68 u.score[3] += (pix1[i] - pix2[i]) * (pix1[i] - pix2[i]);
69
70 return u.score[3];
71 }
72 #endif /* HAVE_ALTIVEC */
73
ff_svq1enc_init_ppc(SVQ1EncContext * c)74 av_cold void ff_svq1enc_init_ppc(SVQ1EncContext *c)
75 {
76 #if HAVE_ALTIVEC
77 if (!PPC_ALTIVEC(av_get_cpu_flags()))
78 return;
79
80 c->ssd_int8_vs_int16 = ssd_int8_vs_int16_altivec;
81 #endif /* HAVE_ALTIVEC */
82 }
83