bdec25223df83fe8f0374bd0bf5ca8f4e2b0d3bd
[deb_ffmpeg.git] / lossless_audiodsp_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 #include "config.h"
22 #if HAVE_ALTIVEC_H
23 #include <altivec.h>
24 #endif
25
26 #include "libavutil/attributes.h"
27 #include "libavutil/cpu.h"
28 #include "libavutil/ppc/cpu.h"
29 #include "libavutil/ppc/types_altivec.h"
30 #include "libavcodec/lossless_audiodsp.h"
31
32 #if HAVE_BIGENDIAN
33 #define GET_T(tt0,tt1,src,a,b){ \
34 a = vec_ld(16, src); \
35 tt0 = vec_perm(b, a, align); \
36 b = vec_ld(32, src); \
37 tt1 = vec_perm(a, b, align); \
38 }
39 #else
40 #define GET_T(tt0,tt1,src,a,b){ \
41 tt0 = vec_vsx_ld(0, src); \
42 tt1 = vec_vsx_ld(16, src); \
43 }
44 #endif
45
46 #if HAVE_ALTIVEC
47 static int32_t scalarproduct_and_madd_int16_altivec(int16_t *v1,
48 const int16_t *v2,
49 const int16_t *v3,
50 int order, int mul)
51 {
52 LOAD_ZERO;
53 vec_s16 *pv1 = (vec_s16 *) v1;
54 register vec_s16 muls = { mul, mul, mul, mul, mul, mul, mul, mul };
55 register vec_s16 t0, t1, i0, i1, i4, i2, i3;
56 register vec_s32 res = zero_s32v;
57 #if HAVE_BIGENDIAN
58 register vec_u8 align = vec_lvsl(0, v2);
59 i2 = vec_ld(0, v2);
60 i3 = vec_ld(0, v3);
61 #endif
62 int32_t ires;
63
64 order >>= 4;
65 do {
66 GET_T(t0,t1,v2,i1,i2);
67 i0 = pv1[0];
68 i1 = pv1[1];
69 res = vec_msum(t0, i0, res);
70 res = vec_msum(t1, i1, res);
71 GET_T(t0,t1,v3,i4,i3);
72 pv1[0] = vec_mladd(t0, muls, i0);
73 pv1[1] = vec_mladd(t1, muls, i1);
74 pv1 += 2;
75 v2 += 16;
76 v3 += 16;
77 } while (--order);
78 res = vec_splat(vec_sums(res, zero_s32v), 3);
79 vec_ste(res, 0, &ires);
80
81 return ires;
82 }
83 #endif /* HAVE_ALTIVEC */
84
85 av_cold void ff_llauddsp_init_ppc(LLAudDSPContext *c)
86 {
87 #if HAVE_ALTIVEC
88 if (!PPC_ALTIVEC(av_get_cpu_flags()))
89 return;
90
91 c->scalarproduct_and_madd_int16 = scalarproduct_and_madd_int16_altivec;
92 #endif /* HAVE_ALTIVEC */
93 }