summaryrefslogtreecommitdiff
path: root/libavcodec/ppc/h264qpel.c
blob: aae6725973994ad36ca7506519235d3772f92de3 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
/*
 * Copyright (c) 2004 Romain Dolbeau <romain@dolbeau.org>
 *
 * This file is part of Libav.
 *
 * Libav is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * Libav is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with Libav; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "config.h"
#include "libavutil/attributes.h"
#include "libavutil/cpu.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/ppc/types_altivec.h"
#include "libavutil/ppc/util_altivec.h"
#include "libavcodec/h264qpel.h"
#include "hpeldsp_altivec.h"

#if HAVE_ALTIVEC

#define PUT_OP_U8_ALTIVEC(d, s, dst) d = s
#define AVG_OP_U8_ALTIVEC(d, s, dst) d = vec_avg(dst, s)

#define OP_U8_ALTIVEC                          PUT_OP_U8_ALTIVEC
#define PREFIX_h264_qpel16_h_lowpass_altivec   put_h264_qpel16_h_lowpass_altivec
#define PREFIX_h264_qpel16_h_lowpass_num       altivec_put_h264_qpel16_h_lowpass_num
#define PREFIX_h264_qpel16_v_lowpass_altivec   put_h264_qpel16_v_lowpass_altivec
#define PREFIX_h264_qpel16_v_lowpass_num       altivec_put_h264_qpel16_v_lowpass_num
#define PREFIX_h264_qpel16_hv_lowpass_altivec  put_h264_qpel16_hv_lowpass_altivec
#define PREFIX_h264_qpel16_hv_lowpass_num      altivec_put_h264_qpel16_hv_lowpass_num
#include "h264qpel_template.c"
#undef OP_U8_ALTIVEC
#undef PREFIX_h264_qpel16_h_lowpass_altivec
#undef PREFIX_h264_qpel16_h_lowpass_num
#undef PREFIX_h264_qpel16_v_lowpass_altivec
#undef PREFIX_h264_qpel16_v_lowpass_num
#undef PREFIX_h264_qpel16_hv_lowpass_altivec
#undef PREFIX_h264_qpel16_hv_lowpass_num

#define OP_U8_ALTIVEC                          AVG_OP_U8_ALTIVEC
#define PREFIX_h264_qpel16_h_lowpass_altivec   avg_h264_qpel16_h_lowpass_altivec
#define PREFIX_h264_qpel16_h_lowpass_num       altivec_avg_h264_qpel16_h_lowpass_num
#define PREFIX_h264_qpel16_v_lowpass_altivec   avg_h264_qpel16_v_lowpass_altivec
#define PREFIX_h264_qpel16_v_lowpass_num       altivec_avg_h264_qpel16_v_lowpass_num
#define PREFIX_h264_qpel16_hv_lowpass_altivec  avg_h264_qpel16_hv_lowpass_altivec
#define PREFIX_h264_qpel16_hv_lowpass_num      altivec_avg_h264_qpel16_hv_lowpass_num
#include "h264qpel_template.c"
#undef OP_U8_ALTIVEC
#undef PREFIX_h264_qpel16_h_lowpass_altivec
#undef PREFIX_h264_qpel16_h_lowpass_num
#undef PREFIX_h264_qpel16_v_lowpass_altivec
#undef PREFIX_h264_qpel16_v_lowpass_num
#undef PREFIX_h264_qpel16_hv_lowpass_altivec
#undef PREFIX_h264_qpel16_hv_lowpass_num

#define H264_MC(OPNAME, SIZE, CODETYPE) \
static void OPNAME ## h264_qpel ## SIZE ## _mc00_ ## CODETYPE (uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    ff_ ## OPNAME ## pixels ## SIZE ## _ ## CODETYPE(dst, src, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc10_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{ \
    DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
    put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src, half, stride, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc20_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    OPNAME ## h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(dst, src, stride, stride);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc30_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
    put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src+1, half, stride, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc01_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
    put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src, half, stride, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc02_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    OPNAME ## h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(dst, src, stride, stride);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc03_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, half)[SIZE*SIZE];\
    put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(half, src, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src+stride, half, stride, stride, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc11_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
    put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\
    put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc31_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
    put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\
    put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src+1, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc13_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
    put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\
    put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc33_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
    put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\
    put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src+1, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc22_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
    OPNAME ## h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(dst, tmp, src, stride, SIZE, stride);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc21_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
    put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\
    put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfHV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc23_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, halfH)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
    put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\
    put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfHV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc12_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
    put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src, SIZE, stride);\
    put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfV, halfHV, stride, SIZE, SIZE);\
}\
\
static void OPNAME ## h264_qpel ## SIZE ## _mc32_ ## CODETYPE(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
{\
    DECLARE_ALIGNED(16, uint8_t, halfV)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, uint8_t, halfHV)[SIZE*SIZE];\
    DECLARE_ALIGNED(16, int16_t, tmp)[SIZE*(SIZE+8)];\
    put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src+1, SIZE, stride);\
    put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\
    OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfV, halfHV, stride, SIZE, SIZE);\
}\

static inline void put_pixels16_l2_altivec( uint8_t * dst, const uint8_t * src1,
                                    const uint8_t * src2, int dst_stride,
                                    int src_stride1, int h)
{
    int i;
    vec_u8 a, b, d, tmp1, tmp2, mask, mask_, edges, align;

    mask_ = vec_lvsl(0, src2);

    for (i = 0; i < h; i++) {

        tmp1 = vec_ld(i * src_stride1, src1);
        mask = vec_lvsl(i * src_stride1, src1);
        tmp2 = vec_ld(i * src_stride1 + 15, src1);

        a = vec_perm(tmp1, tmp2, mask);

        tmp1 = vec_ld(i * 16, src2);
        tmp2 = vec_ld(i * 16 + 15, src2);

        b = vec_perm(tmp1, tmp2, mask_);

        tmp1 = vec_ld(0, dst);
        mask = vec_lvsl(0, dst);
        tmp2 = vec_ld(15, dst);

        d = vec_avg(a, b);

        edges = vec_perm(tmp2, tmp1, mask);

        align = vec_lvsr(0, dst);

        tmp2 = vec_perm(d, edges, align);
        tmp1 = vec_perm(edges, d, align);

        vec_st(tmp2, 15, dst);
        vec_st(tmp1, 0 , dst);

        dst += dst_stride;
    }
}

static inline void avg_pixels16_l2_altivec( uint8_t * dst, const uint8_t * src1,
                                    const uint8_t * src2, int dst_stride,
                                    int src_stride1, int h)
{
    int i;
    vec_u8 a, b, d, tmp1, tmp2, mask, mask_, edges, align;

    mask_ = vec_lvsl(0, src2);

    for (i = 0; i < h; i++) {

        tmp1 = vec_ld(i * src_stride1, src1);
        mask = vec_lvsl(i * src_stride1, src1);
        tmp2 = vec_ld(i * src_stride1 + 15, src1);

        a = vec_perm(tmp1, tmp2, mask);

        tmp1 = vec_ld(i * 16, src2);
        tmp2 = vec_ld(i * 16 + 15, src2);

        b = vec_perm(tmp1, tmp2, mask_);

        tmp1 = vec_ld(0, dst);
        mask = vec_lvsl(0, dst);
        tmp2 = vec_ld(15, dst);

        d = vec_avg(vec_perm(tmp1, tmp2, mask), vec_avg(a, b));

        edges = vec_perm(tmp2, tmp1, mask);

        align = vec_lvsr(0, dst);

        tmp2 = vec_perm(d, edges, align);
        tmp1 = vec_perm(edges, d, align);

        vec_st(tmp2, 15, dst);
        vec_st(tmp1, 0 , dst);

        dst += dst_stride;
    }
}

/* Implemented but could be faster
#define put_pixels16_l2_altivec(d,s1,s2,ds,s1s,h) put_pixels16_l2(d,s1,s2,ds,s1s,16,h)
#define avg_pixels16_l2_altivec(d,s1,s2,ds,s1s,h) avg_pixels16_l2(d,s1,s2,ds,s1s,16,h)
 */

H264_MC(put_, 16, altivec)
H264_MC(avg_, 16, altivec)
#endif /* HAVE_ALTIVEC */

av_cold void ff_h264qpel_init_ppc(H264QpelContext *c, int bit_depth)
{
#if HAVE_ALTIVEC
    const int high_bit_depth = bit_depth > 8;

    if (!(av_get_cpu_flags() & AV_CPU_FLAG_ALTIVEC))
        return;

    if (!high_bit_depth) {
#define dspfunc(PFX, IDX, NUM) \
        c->PFX ## _pixels_tab[IDX][ 0] = PFX ## NUM ## _mc00_altivec; \
        c->PFX ## _pixels_tab[IDX][ 1] = PFX ## NUM ## _mc10_altivec; \
        c->PFX ## _pixels_tab[IDX][ 2] = PFX ## NUM ## _mc20_altivec; \
        c->PFX ## _pixels_tab[IDX][ 3] = PFX ## NUM ## _mc30_altivec; \
        c->PFX ## _pixels_tab[IDX][ 4] = PFX ## NUM ## _mc01_altivec; \
        c->PFX ## _pixels_tab[IDX][ 5] = PFX ## NUM ## _mc11_altivec; \
        c->PFX ## _pixels_tab[IDX][ 6] = PFX ## NUM ## _mc21_altivec; \
        c->PFX ## _pixels_tab[IDX][ 7] = PFX ## NUM ## _mc31_altivec; \
        c->PFX ## _pixels_tab[IDX][ 8] = PFX ## NUM ## _mc02_altivec; \
        c->PFX ## _pixels_tab[IDX][ 9] = PFX ## NUM ## _mc12_altivec; \
        c->PFX ## _pixels_tab[IDX][10] = PFX ## NUM ## _mc22_altivec; \
        c->PFX ## _pixels_tab[IDX][11] = PFX ## NUM ## _mc32_altivec; \
        c->PFX ## _pixels_tab[IDX][12] = PFX ## NUM ## _mc03_altivec; \
        c->PFX ## _pixels_tab[IDX][13] = PFX ## NUM ## _mc13_altivec; \
        c->PFX ## _pixels_tab[IDX][14] = PFX ## NUM ## _mc23_altivec; \
        c->PFX ## _pixels_tab[IDX][15] = PFX ## NUM ## _mc33_altivec

        dspfunc(put_h264_qpel, 0, 16);
        dspfunc(avg_h264_qpel, 0, 16);
#undef dspfunc
    }
#endif /* HAVE_ALTIVEC */
}