1/*
2 * Copyright (c) 2006 Luca Barbato <lu_zero@gentoo.org>
3 *
4 * This file is part of Libav.
5 *
6 * Libav 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 * Libav 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 Libav; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19 */
20
21#include "libavcodec/dsputil.h"
22
23#include "dsputil_altivec.h"
24#include "util_altivec.h"
25
26static void vector_fmul_altivec(float *dst, const float *src0, const float *src1, int len)
27{
28    int i;
29    vector float d0, d1, s, zero = (vector float)vec_splat_u32(0);
30    for(i=0; i<len-7; i+=8) {
31        d0 = vec_ld(0, src0+i);
32        s = vec_ld(0, src1+i);
33        d1 = vec_ld(16, src0+i);
34        d0 = vec_madd(d0, s, zero);
35        d1 = vec_madd(d1, vec_ld(16,src1+i), zero);
36        vec_st(d0, 0, dst+i);
37        vec_st(d1, 16, dst+i);
38    }
39}
40
41static void vector_fmul_reverse_altivec(float *dst, const float *src0,
42                                        const float *src1, int len)
43{
44    int i;
45    vector float d, s0, s1, h0, l0,
46                 s2, s3, zero = (vector float)vec_splat_u32(0);
47    src1 += len-4;
48    for(i=0; i<len-7; i+=8) {
49        s1 = vec_ld(0, src1-i);              // [a,b,c,d]
50        s0 = vec_ld(0, src0+i);
51        l0 = vec_mergel(s1, s1);             // [c,c,d,d]
52        s3 = vec_ld(-16, src1-i);
53        h0 = vec_mergeh(s1, s1);             // [a,a,b,b]
54        s2 = vec_ld(16, src0+i);
55        s1 = vec_mergeh(vec_mergel(l0,h0),   // [d,b,d,b]
56                        vec_mergeh(l0,h0));  // [c,a,c,a]
57                                             // [d,c,b,a]
58        l0 = vec_mergel(s3, s3);
59        d = vec_madd(s0, s1, zero);
60        h0 = vec_mergeh(s3, s3);
61        vec_st(d, 0, dst+i);
62        s3 = vec_mergeh(vec_mergel(l0,h0),
63                        vec_mergeh(l0,h0));
64        d = vec_madd(s2, s3, zero);
65        vec_st(d, 16, dst+i);
66    }
67}
68
69static void vector_fmul_add_altivec(float *dst, const float *src0,
70                                    const float *src1, const float *src2,
71                                    int len)
72{
73    int i;
74    vector float d, s0, s1, s2, t0, t1, edges;
75    vector unsigned char align = vec_lvsr(0,dst),
76                         mask = vec_lvsl(0, dst);
77
78    for (i=0; i<len-3; i+=4) {
79        t0 = vec_ld(0, dst+i);
80        t1 = vec_ld(15, dst+i);
81        s0 = vec_ld(0, src0+i);
82        s1 = vec_ld(0, src1+i);
83        s2 = vec_ld(0, src2+i);
84        edges = vec_perm(t1 ,t0, mask);
85        d = vec_madd(s0,s1,s2);
86        t1 = vec_perm(d, edges, align);
87        t0 = vec_perm(edges, d, align);
88        vec_st(t1, 15, dst+i);
89        vec_st(t0, 0, dst+i);
90    }
91}
92
93static void vector_fmul_window_altivec(float *dst, const float *src0, const float *src1, const float *win, int len)
94{
95    vector float zero, t0, t1, s0, s1, wi, wj;
96    const vector unsigned char reverse = vcprm(3,2,1,0);
97    int i,j;
98
99    dst += len;
100    win += len;
101    src0+= len;
102
103    zero = (vector float)vec_splat_u32(0);
104
105    for(i=-len*4, j=len*4-16; i<0; i+=16, j-=16) {
106        s0 = vec_ld(i, src0);
107        s1 = vec_ld(j, src1);
108        wi = vec_ld(i, win);
109        wj = vec_ld(j, win);
110
111        s1 = vec_perm(s1, s1, reverse);
112        wj = vec_perm(wj, wj, reverse);
113
114        t0 = vec_madd(s0, wj, zero);
115        t0 = vec_nmsub(s1, wi, t0);
116        t1 = vec_madd(s0, wi, zero);
117        t1 = vec_madd(s1, wj, t1);
118        t1 = vec_perm(t1, t1, reverse);
119
120        vec_st(t0, i, dst);
121        vec_st(t1, j, dst);
122    }
123}
124
125void float_init_altivec(DSPContext* c, AVCodecContext *avctx)
126{
127    c->vector_fmul = vector_fmul_altivec;
128    c->vector_fmul_reverse = vector_fmul_reverse_altivec;
129    c->vector_fmul_add = vector_fmul_add_altivec;
130    if(!(avctx->flags & CODEC_FLAG_BITEXACT)) {
131        c->vector_fmul_window = vector_fmul_window_altivec;
132    }
133}
134