summaryrefslogtreecommitdiffstats
path: root/third_party/aom/aom_dsp/mips/convolve8_dspr2.c
blob: af54b4264777a05a4e1e89e96f86b12707140aec (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
/*
 * Copyright (c) 2016, Alliance for Open Media. All rights reserved
 *
 * This source code is subject to the terms of the BSD 2 Clause License and
 * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
 * was not distributed with this source code in the LICENSE file, you can
 * obtain it at www.aomedia.org/license/software. If the Alliance for Open
 * Media Patent License 1.0 was not distributed with this source code in the
 * PATENTS file, you can obtain it at www.aomedia.org/license/patent.
 */

#include <assert.h>
#include <stdio.h>

#include "config/aom_dsp_rtcd.h"

#include "aom_dsp/mips/convolve_common_dspr2.h"
#include "aom_dsp/aom_dsp_common.h"
#include "aom_dsp/aom_filter.h"
#include "aom_ports/mem.h"

#if HAVE_DSPR2
void aom_convolve_copy_dspr2(const uint8_t *src, ptrdiff_t src_stride,
                             uint8_t *dst, ptrdiff_t dst_stride,
                             const int16_t *filter_x, int filter_x_stride,
                             const int16_t *filter_y, int filter_y_stride,
                             int w, int h) {
  int x, y;

  (void)filter_x;
  (void)filter_x_stride;
  (void)filter_y;
  (void)filter_y_stride;

  /* prefetch data to cache memory */
  prefetch_load(src);
  prefetch_load(src + 32);
  prefetch_store(dst);

  switch (w) {
    case 4: {
      uint32_t tp1;

      /* 1 word storage */
      for (y = h; y--;) {
        prefetch_load(src + src_stride);
        prefetch_load(src + src_stride + 32);
        prefetch_store(dst + dst_stride);

        __asm__ __volatile__(
            "ulw              %[tp1],         (%[src])      \n\t"
            "sw               %[tp1],         (%[dst])      \n\t" /* store */

            : [tp1] "=&r"(tp1)
            : [src] "r"(src), [dst] "r"(dst));

        src += src_stride;
        dst += dst_stride;
      }
    } break;
    case 8: {
      uint32_t tp1, tp2;

      /* 2 word storage */
      for (y = h; y--;) {
        prefetch_load(src + src_stride);
        prefetch_load(src + src_stride + 32);
        prefetch_store(dst + dst_stride);

        __asm__ __volatile__(
            "ulw              %[tp1],         0(%[src])      \n\t"
            "ulw              %[tp2],         4(%[src])      \n\t"
            "sw               %[tp1],         0(%[dst])      \n\t" /* store */
            "sw               %[tp2],         4(%[dst])      \n\t" /* store */

            : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2)
            : [src] "r"(src), [dst] "r"(dst));

        src += src_stride;
        dst += dst_stride;
      }
    } break;
    case 16: {
      uint32_t tp1, tp2, tp3, tp4;

      /* 4 word storage */
      for (y = h; y--;) {
        prefetch_load(src + src_stride);
        prefetch_load(src + src_stride + 32);
        prefetch_store(dst + dst_stride);

        __asm__ __volatile__(
            "ulw              %[tp1],         0(%[src])      \n\t"
            "ulw              %[tp2],         4(%[src])      \n\t"
            "ulw              %[tp3],         8(%[src])      \n\t"
            "ulw              %[tp4],         12(%[src])     \n\t"

            "sw               %[tp1],         0(%[dst])      \n\t" /* store */
            "sw               %[tp2],         4(%[dst])      \n\t" /* store */
            "sw               %[tp3],         8(%[dst])      \n\t" /* store */
            "sw               %[tp4],         12(%[dst])     \n\t" /* store */

            : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2), [tp3] "=&r"(tp3),
              [tp4] "=&r"(tp4)
            : [src] "r"(src), [dst] "r"(dst));

        src += src_stride;
        dst += dst_stride;
      }
    } break;
    case 32: {
      uint32_t tp1, tp2, tp3, tp4;
      uint32_t tp5, tp6, tp7, tp8;

      /* 8 word storage */
      for (y = h; y--;) {
        prefetch_load(src + src_stride);
        prefetch_load(src + src_stride + 32);
        prefetch_store(dst + dst_stride);

        __asm__ __volatile__(
            "ulw              %[tp1],         0(%[src])      \n\t"
            "ulw              %[tp2],         4(%[src])      \n\t"
            "ulw              %[tp3],         8(%[src])      \n\t"
            "ulw              %[tp4],         12(%[src])     \n\t"
            "ulw              %[tp5],         16(%[src])     \n\t"
            "ulw              %[tp6],         20(%[src])     \n\t"
            "ulw              %[tp7],         24(%[src])     \n\t"
            "ulw              %[tp8],         28(%[src])     \n\t"

            "sw               %[tp1],         0(%[dst])      \n\t" /* store */
            "sw               %[tp2],         4(%[dst])      \n\t" /* store */
            "sw               %[tp3],         8(%[dst])      \n\t" /* store */
            "sw               %[tp4],         12(%[dst])     \n\t" /* store */
            "sw               %[tp5],         16(%[dst])     \n\t" /* store */
            "sw               %[tp6],         20(%[dst])     \n\t" /* store */
            "sw               %[tp7],         24(%[dst])     \n\t" /* store */
            "sw               %[tp8],         28(%[dst])     \n\t" /* store */

            : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2), [tp3] "=&r"(tp3),
              [tp4] "=&r"(tp4), [tp5] "=&r"(tp5), [tp6] "=&r"(tp6),
              [tp7] "=&r"(tp7), [tp8] "=&r"(tp8)
            : [src] "r"(src), [dst] "r"(dst));

        src += src_stride;
        dst += dst_stride;
      }
    } break;
    case 64: {
      uint32_t tp1, tp2, tp3, tp4;
      uint32_t tp5, tp6, tp7, tp8;

      prefetch_load(src + 64);
      prefetch_store(dst + 32);

      /* 16 word storage */
      for (y = h; y--;) {
        prefetch_load(src + src_stride);
        prefetch_load(src + src_stride + 32);
        prefetch_load(src + src_stride + 64);
        prefetch_store(dst + dst_stride);
        prefetch_store(dst + dst_stride + 32);

        __asm__ __volatile__(
            "ulw              %[tp1],         0(%[src])      \n\t"
            "ulw              %[tp2],         4(%[src])      \n\t"
            "ulw              %[tp3],         8(%[src])      \n\t"
            "ulw              %[tp4],         12(%[src])     \n\t"
            "ulw              %[tp5],         16(%[src])     \n\t"
            "ulw              %[tp6],         20(%[src])     \n\t"
            "ulw              %[tp7],         24(%[src])     \n\t"
            "ulw              %[tp8],         28(%[src])     \n\t"

            "sw               %[tp1],         0(%[dst])      \n\t" /* store */
            "sw               %[tp2],         4(%[dst])      \n\t" /* store */
            "sw               %[tp3],         8(%[dst])      \n\t" /* store */
            "sw               %[tp4],         12(%[dst])     \n\t" /* store */
            "sw               %[tp5],         16(%[dst])     \n\t" /* store */
            "sw               %[tp6],         20(%[dst])     \n\t" /* store */
            "sw               %[tp7],         24(%[dst])     \n\t" /* store */
            "sw               %[tp8],         28(%[dst])     \n\t" /* store */

            "ulw              %[tp1],         32(%[src])     \n\t"
            "ulw              %[tp2],         36(%[src])     \n\t"
            "ulw              %[tp3],         40(%[src])     \n\t"
            "ulw              %[tp4],         44(%[src])     \n\t"
            "ulw              %[tp5],         48(%[src])     \n\t"
            "ulw              %[tp6],         52(%[src])     \n\t"
            "ulw              %[tp7],         56(%[src])     \n\t"
            "ulw              %[tp8],         60(%[src])     \n\t"

            "sw               %[tp1],         32(%[dst])     \n\t" /* store */
            "sw               %[tp2],         36(%[dst])     \n\t" /* store */
            "sw               %[tp3],         40(%[dst])     \n\t" /* store */
            "sw               %[tp4],         44(%[dst])     \n\t" /* store */
            "sw               %[tp5],         48(%[dst])     \n\t" /* store */
            "sw               %[tp6],         52(%[dst])     \n\t" /* store */
            "sw               %[tp7],         56(%[dst])     \n\t" /* store */
            "sw               %[tp8],         60(%[dst])     \n\t" /* store */

            : [tp1] "=&r"(tp1), [tp2] "=&r"(tp2), [tp3] "=&r"(tp3),
              [tp4] "=&r"(tp4), [tp5] "=&r"(tp5), [tp6] "=&r"(tp6),
              [tp7] "=&r"(tp7), [tp8] "=&r"(tp8)
            : [src] "r"(src), [dst] "r"(dst));

        src += src_stride;
        dst += dst_stride;
      }
    } break;
    default:
      for (y = h; y--;) {
        for (x = 0; x < w; ++x) {
          dst[x] = src[x];
        }

        src += src_stride;
        dst += dst_stride;
      }
      break;
  }
}
#endif