summaryrefslogtreecommitdiffstats
path: root/libavcodec/ppc/pixblockdsp.c
blob: 84aa562bb67fcc065480a4e2e750bf7c904b218d (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
/*
 * Copyright (c) 2002 Brian Foley
 * Copyright (c) 2002 Dieter Shirley
 * Copyright (c) 2003-2004 Romain Dolbeau <romain@dolbeau.org>
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg 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.
 *
 * FFmpeg 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 FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "config.h"
#if HAVE_ALTIVEC_H
#include <altivec.h>
#endif

#include "libavutil/attributes.h"
#include "libavutil/cpu.h"
#include "libavutil/ppc/cpu.h"
#include "libavutil/ppc/types_altivec.h"
#include "libavutil/ppc/util_altivec.h"
#include "libavcodec/avcodec.h"
#include "libavcodec/pixblockdsp.h"

#if HAVE_ALTIVEC

#if HAVE_VSX
static void get_pixels_altivec(int16_t *restrict block, const uint8_t *pixels,
                               ptrdiff_t line_size)
{
    int i;
    vector unsigned char perm =
        (vector unsigned char) {0x00,0x10, 0x01,0x11,0x02,0x12,0x03,0x13,\
            0x04,0x14,0x05,0x15,0x06,0x16,0x07,0x17};
    const vector unsigned char zero =
        (const vector unsigned char) vec_splat_u8(0);

    for (i = 0; i < 8; i++) {
        /* Read potentially unaligned pixels.
         * We're reading 16 pixels, and actually only want 8,
         * but we simply ignore the extras. */
        vector unsigned char bytes = vec_vsx_ld(0, pixels);

        // Convert the bytes into shorts.
        //vector signed short shorts = (vector signed short) vec_perm(zero, bytes, perm);
        vector signed short shorts = (vector signed short) vec_perm(bytes, zero, perm);

        // Save the data to the block, we assume the block is 16-byte aligned.
        vec_vsx_st(shorts, i * 16, (vector signed short *) block);

        pixels += line_size;
    }
}
#else
static void get_pixels_altivec(int16_t *restrict block, const uint8_t *pixels,
                               ptrdiff_t line_size)
{
    int i;
    vec_u8 perm = vec_lvsl(0, pixels);
    const vec_u8 zero = (const vec_u8)vec_splat_u8(0);

    for (i = 0; i < 8; i++) {
        /* Read potentially unaligned pixels.
         * We're reading 16 pixels, and actually only want 8,
         * but we simply ignore the extras. */
        vec_u8 pixl = vec_ld(0, pixels);
        vec_u8 pixr = vec_ld(7, pixels);
        vec_u8 bytes = vec_perm(pixl, pixr, perm);

        // Convert the bytes into shorts.
        vec_s16 shorts = (vec_s16)vec_mergeh(zero, bytes);

        // Save the data to the block, we assume the block is 16-byte aligned.
        vec_st(shorts, i * 16, (vec_s16 *)block);

        pixels += line_size;
    }
}

#endif /* HAVE_VSX */

#if HAVE_VSX
static void diff_pixels_altivec(int16_t *restrict block, const uint8_t *s1,
                                const uint8_t *s2, int stride)
{
  int i;
  const vector unsigned char zero =
    (const vector unsigned char) vec_splat_u8(0);
  vector signed short shorts1, shorts2;

  for (i = 0; i < 4; i++) {
    /* Read potentially unaligned pixels.
     * We're reading 16 pixels, and actually only want 8,
     * but we simply ignore the extras. */
    vector unsigned char bytes = vec_vsx_ld(0,  s1);

    // Convert the bytes into shorts.
    shorts1 = (vector signed short) vec_mergeh(bytes, zero);

    // Do the same for the second block of pixels.
    bytes =vec_vsx_ld(0,  s2);

    // Convert the bytes into shorts.
    shorts2 = (vector signed short) vec_mergeh(bytes, zero);

    // Do the subtraction.
    shorts1 = vec_sub(shorts1, shorts2);

    // Save the data to the block, we assume the block is 16-byte aligned.
    vec_vsx_st(shorts1, 0, (vector signed short *) block);

    s1    += stride;
    s2    += stride;
    block += 8;

    /* The code below is a copy of the code above...
     * This is a manual unroll. */

    /* Read potentially unaligned pixels.
     * We're reading 16 pixels, and actually only want 8,
     * but we simply ignore the extras. */
    bytes = vec_vsx_ld(0,  s1);

    // Convert the bytes into shorts.
    shorts1 = (vector signed short) vec_mergeh(bytes, zero);

    // Do the same for the second block of pixels.
    bytes = vec_vsx_ld(0,  s2);

    // Convert the bytes into shorts.
    shorts2 = (vector signed short) vec_mergeh(bytes, zero);

    // Do the subtraction.
    shorts1 = vec_sub(shorts1, shorts2);

    // Save the data to the block, we assume the block is 16-byte aligned.
    vec_vsx_st(shorts1, 0, (vector signed short *) block);

    s1    += stride;
    s2    += stride;
    block += 8;
  }
}
#else
static void diff_pixels_altivec(int16_t *restrict block, const uint8_t *s1,
                                const uint8_t *s2, int stride)
{
    int i;
    vec_u8 perm1 = vec_lvsl(0, s1);
    vec_u8 perm2 = vec_lvsl(0, s2);
    const vec_u8 zero = (const vec_u8)vec_splat_u8(0);
    vec_s16 shorts1, shorts2;

    for (i = 0; i < 4; i++) {
        /* Read potentially unaligned pixels.
         * We're reading 16 pixels, and actually only want 8,
         * but we simply ignore the extras. */
        vec_u8 pixl  = vec_ld(0,  s1);
        vec_u8 pixr  = vec_ld(15, s1);
        vec_u8 bytes = vec_perm(pixl, pixr, perm1);

        // Convert the bytes into shorts.
        shorts1 = (vec_s16)vec_mergeh(zero, bytes);

        // Do the same for the second block of pixels.
        pixl  = vec_ld(0,  s2);
        pixr  = vec_ld(15, s2);
        bytes = vec_perm(pixl, pixr, perm2);

        // Convert the bytes into shorts.
        shorts2 = (vec_s16)vec_mergeh(zero, bytes);

        // Do the subtraction.
        shorts1 = vec_sub(shorts1, shorts2);

        // Save the data to the block, we assume the block is 16-byte aligned.
        vec_st(shorts1, 0, (vec_s16 *)block);

        s1    += stride;
        s2    += stride;
        block += 8;

        /* The code below is a copy of the code above...
         * This is a manual unroll. */

        /* Read potentially unaligned pixels.
         * We're reading 16 pixels, and actually only want 8,
         * but we simply ignore the extras. */
        pixl  = vec_ld(0,  s1);
        pixr  = vec_ld(15, s1);
        bytes = vec_perm(pixl, pixr, perm1);

        // Convert the bytes into shorts.
        shorts1 = (vec_s16)vec_mergeh(zero, bytes);

        // Do the same for the second block of pixels.
        pixl  = vec_ld(0,  s2);
        pixr  = vec_ld(15, s2);
        bytes = vec_perm(pixl, pixr, perm2);

        // Convert the bytes into shorts.
        shorts2 = (vec_s16)vec_mergeh(zero, bytes);

        // Do the subtraction.
        shorts1 = vec_sub(shorts1, shorts2);

        // Save the data to the block, we assume the block is 16-byte aligned.
        vec_st(shorts1, 0, (vec_s16 *)block);

        s1    += stride;
        s2    += stride;
        block += 8;
    }
}

#endif /* HAVE_VSX */

#endif /* HAVE_ALTIVEC */

#if HAVE_VSX
static void get_pixels_vsx(int16_t *restrict block, const uint8_t *pixels,
                           ptrdiff_t line_size)
{
    int i;
    for (i = 0; i < 8; i++) {
        vec_s16 shorts = vsx_ld_u8_s16(0, pixels);

        vec_vsx_st(shorts, i * 16, block);

        pixels += line_size;
    }
}

static void diff_pixels_vsx(int16_t *restrict block, const uint8_t *s1,
                            const uint8_t *s2, int stride)
{
    int i;
    vec_s16 shorts1, shorts2;
    for (i = 0; i < 8; i++) {
        shorts1 = vsx_ld_u8_s16(0, s1);
        shorts2 = vsx_ld_u8_s16(0, s2);

        shorts1 = vec_sub(shorts1, shorts2);

        vec_vsx_st(shorts1, 0, block);

        s1    += stride;
        s2    += stride;
        block += 8;
    }
}
#endif /* HAVE_VSX */

av_cold void ff_pixblockdsp_init_ppc(PixblockDSPContext *c,
                                     AVCodecContext *avctx,
                                     unsigned high_bit_depth)
{
#if HAVE_ALTIVEC
    if (!PPC_ALTIVEC(av_get_cpu_flags()))
        return;

    c->diff_pixels = diff_pixels_altivec;

    if (!high_bit_depth) {
        c->get_pixels = get_pixels_altivec;
    }
#endif /* HAVE_ALTIVEC */

#if HAVE_VSX
    if (!PPC_VSX(av_get_cpu_flags()))
        return;

    c->diff_pixels = diff_pixels_vsx;

    if (!high_bit_depth)
        c->get_pixels = get_pixels_vsx;
#endif /* HAVE_VSX */
}
OpenPOWER on IntegriCloud