av1_rtcd_defs.pl 52.5 KB
Newer Older
Yaowu Xu's avatar
Yaowu Xu committed
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
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
sub av1_common_forward_decls() {
print <<EOF
/*
 * AV1
 */

#include "aom/aom_integer.h"
#include "av1/common/common.h"
#include "av1/common/enums.h"
#include "av1/common/quant_common.h"
#include "av1/common/filter.h"
#include "av1/common/av1_txfm.h"

struct macroblockd;

/* Encoder forward decls */
struct macroblock;
struct aom_variance_vtable;
struct search_site_config;
struct mv;
union int_mv;
struct yv12_buffer_config;
EOF
}
forward_decls qw/av1_common_forward_decls/;

# functions that are 64 bit only.
$mmx_x86_64 = $sse2_x86_64 = $ssse3_x86_64 = $avx_x86_64 = $avx2_x86_64 = '';
if ($opts{arch} eq "x86_64") {
  $mmx_x86_64 = 'mmx';
  $sse2_x86_64 = 'sse2';
  $ssse3_x86_64 = 'ssse3';
  $avx_x86_64 = 'avx';
  $avx2_x86_64 = 'avx2';
}

#
# 10/12-tap convolution filters
#
add_proto qw/void av1_convolve_horiz/, "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, int w, int h, const InterpFilterParams fp, const int subpel_x_q4, int x_step_q4, int avg";
specialize qw/av1_convolve_horiz ssse3/;

add_proto qw/void av1_convolve_vert/, "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, int w, int h, const InterpFilterParams fp, const int subpel_x_q4, int x_step_q4, int avg";
specialize qw/av1_convolve_vert ssse3/;

if (aom_config("CONFIG_AOM_HIGHBITDEPTH") eq "yes") {
  add_proto qw/void av1_highbd_convolve_horiz/, "const uint16_t *src, int src_stride, uint16_t *dst, int dst_stride, int w, int h, const InterpFilterParams fp, const int subpel_x_q4, int x_step_q4, int avg, int bd";
  specialize qw/av1_highbd_convolve_horiz sse4_1/;
  add_proto qw/void av1_highbd_convolve_vert/, "const uint16_t *src, int src_stride, uint16_t *dst, int dst_stride, int w, int h, const InterpFilterParams fp, const int subpel_x_q4, int x_step_q4, int avg, int bd";
  specialize qw/av1_highbd_convolve_vert sse4_1/;
}

#
# dct
#
if (aom_config("CONFIG_AOM_HIGHBITDEPTH") eq "yes") {
  # Note as optimized versions of these functions are added we need to add a check to ensure
  # that when CONFIG_EMULATE_HARDWARE is on, it defaults to the C versions only.
  if (aom_config("CONFIG_EMULATE_HARDWARE") eq "yes") {
    add_proto qw/void av1_iht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht4x4_16_add/;

    add_proto qw/void av1_iht4x8_32_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht4x8_32_add/;

    add_proto qw/void av1_iht8x4_32_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x4_32_add/;

    add_proto qw/void av1_iht8x16_128_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x16_128_add/;

    add_proto qw/void av1_iht16x8_128_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht16x8_128_add/;

    add_proto qw/void av1_iht16x32_512_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht16x32_512_add/;

    add_proto qw/void av1_iht32x16_512_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht32x16_512_add/;

    add_proto qw/void av1_iht8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x8_64_add/;

    add_proto qw/void av1_iht16x16_256_add/, "const tran_low_t *input, uint8_t *output, int pitch, int tx_type";
    specialize qw/av1_iht16x16_256_add/;
  } else {
    add_proto qw/void av1_iht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht4x4_16_add sse2/;

    add_proto qw/void av1_iht4x8_32_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht4x8_32_add/;

    add_proto qw/void av1_iht8x4_32_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x4_32_add/;

    add_proto qw/void av1_iht8x16_128_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x16_128_add/;

    add_proto qw/void av1_iht16x8_128_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht16x8_128_add/;

    add_proto qw/void av1_iht16x32_512_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht16x32_512_add/;

    add_proto qw/void av1_iht32x16_512_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht32x16_512_add/;

    add_proto qw/void av1_iht8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x8_64_add sse2/;

    add_proto qw/void av1_iht16x16_256_add/, "const tran_low_t *input, uint8_t *output, int pitch, int tx_type";
    specialize qw/av1_iht16x16_256_add sse2/;
  }
} else {
  # Force C versions if CONFIG_EMULATE_HARDWARE is 1
  if (aom_config("CONFIG_EMULATE_HARDWARE") eq "yes") {
    add_proto qw/void av1_iht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht4x4_16_add/;

    add_proto qw/void av1_iht4x8_32_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht4x8_32_add/;

    add_proto qw/void av1_iht8x4_32_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x4_32_add/;

    add_proto qw/void av1_iht8x16_128_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x16_128_add/;

    add_proto qw/void av1_iht16x8_128_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht16x8_128_add/;

    add_proto qw/void av1_iht16x32_512_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht16x32_512_add/;

    add_proto qw/void av1_iht32x16_512_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht32x16_512_add/;

    add_proto qw/void av1_iht8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x8_64_add/;

    add_proto qw/void av1_iht16x16_256_add/, "const tran_low_t *input, uint8_t *output, int pitch, int tx_type";
    specialize qw/av1_iht16x16_256_add/;
  } else {
    add_proto qw/void av1_iht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht4x4_16_add sse2 neon dspr2/;

    add_proto qw/void av1_iht4x8_32_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht4x8_32_add/;

    add_proto qw/void av1_iht8x4_32_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x4_32_add/;

    add_proto qw/void av1_iht8x16_128_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x16_128_add/;

    add_proto qw/void av1_iht16x8_128_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht16x8_128_add/;

    add_proto qw/void av1_iht16x32_512_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht16x32_512_add/;

    add_proto qw/void av1_iht32x16_512_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht32x16_512_add/;

    add_proto qw/void av1_iht8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
    specialize qw/av1_iht8x8_64_add sse2 neon dspr2/;

    add_proto qw/void av1_iht16x16_256_add/, "const tran_low_t *input, uint8_t *output, int pitch, int tx_type";
    specialize qw/av1_iht16x16_256_add sse2 dspr2/;

    if (aom_config("CONFIG_EXT_TX") ne "yes") {
      specialize qw/av1_iht4x4_16_add msa/;
      specialize qw/av1_iht8x8_64_add msa/;
      specialize qw/av1_iht16x16_256_add msa/;
    }
  }
}

if (aom_config("CONFIG_NEW_QUANT") eq "yes") {
  add_proto qw/void quantize_nuq/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, const int16_t *dequant_ptr, const cuml_bins_type_nuq *cuml_bins_ptr, const dequant_val_type_nuq *dequant_val, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, uint16_t *eob_ptr, const int16_t *scan, const uint8_t *band";
  specialize qw/quantize_nuq/;

  add_proto qw/void quantize_fp_nuq/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *quant_ptr, const int16_t *dequant_ptr, const cuml_bins_type_nuq *cuml_bins_ptr, const dequant_val_type_nuq *dequant_val, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, uint16_t *eob_ptr, const int16_t *scan, const uint8_t *band";
  specialize qw/quantize_fp_nuq/;

  add_proto qw/void quantize_32x32_nuq/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, const int16_t *dequant_ptr, const cuml_bins_type_nuq *cuml_bins_ptr, const dequant_val_type_nuq *dequant_val, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, uint16_t *eob_ptr, const int16_t *scan, const uint8_t *band";
  specialize qw/quantize_32x32_nuq/;

  add_proto qw/void quantize_32x32_fp_nuq/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *quant_ptr, const int16_t *dequant_ptr, const cuml_bins_type_nuq *cuml_bins_ptr, const dequant_val_type_nuq *dequant_val, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, uint16_t *eob_ptr, const int16_t *scan, const uint8_t *band";
  specialize qw/quantize_32x32_fp_nuq/;
}

# EXT_INTRA predictor functions
if (aom_config("CONFIG_EXT_INTRA") eq "yes") {
  add_proto qw/void av1_dc_filter_predictor/, "uint8_t *dst, ptrdiff_t stride, int bs, const uint8_t *above, const uint8_t *left";
  specialize qw/av1_dc_filter_predictor sse4_1/;
  add_proto qw/void av1_v_filter_predictor/, "uint8_t *dst, ptrdiff_t stride, int bs, const uint8_t *above, const uint8_t *left";
  specialize qw/av1_v_filter_predictor sse4_1/;
  add_proto qw/void av1_h_filter_predictor/, "uint8_t *dst, ptrdiff_t stride, int bs, const uint8_t *above, const uint8_t *left";
  specialize qw/av1_h_filter_predictor sse4_1/;
  add_proto qw/void av1_d45_filter_predictor/, "uint8_t *dst, ptrdiff_t stride, int bs, const uint8_t *above, const uint8_t *left";
  specialize qw/av1_d45_filter_predictor sse4_1/;
  add_proto qw/void av1_d135_filter_predictor/, "uint8_t *dst, ptrdiff_t stride, int bs, const uint8_t *above, const uint8_t *left";
  specialize qw/av1_d135_filter_predictor sse4_1/;
  add_proto qw/void av1_d117_filter_predictor/, "uint8_t *dst, ptrdiff_t stride, int bs, const uint8_t *above, const uint8_t *left";
  specialize qw/av1_d117_filter_predictor sse4_1/;
  add_proto qw/void av1_d153_filter_predictor/, "uint8_t *dst, ptrdiff_t stride, int bs, const uint8_t *above, const uint8_t *left";
  specialize qw/av1_d153_filter_predictor sse4_1/;
  add_proto qw/void av1_d207_filter_predictor/, "uint8_t *dst, ptrdiff_t stride, int bs, const uint8_t *above, const uint8_t *left";
  specialize qw/av1_d207_filter_predictor sse4_1/;
  add_proto qw/void av1_d63_filter_predictor/, "uint8_t *dst, ptrdiff_t stride, int bs, const uint8_t *above, const uint8_t *left";
  specialize qw/av1_d63_filter_predictor sse4_1/;
  add_proto qw/void av1_tm_filter_predictor/, "uint8_t *dst, ptrdiff_t stride, int bs, const uint8_t *above, const uint8_t *left";
  specialize qw/av1_tm_filter_predictor sse4_1/;
  # High bitdepth functions
  if (aom_config("CONFIG_AOM_HIGHBITDEPTH") eq "yes") {
    add_proto qw/void av1_highbd_dc_filter_predictor/, "uint16_t *dst, ptrdiff_t stride, int bs, const uint16_t *above, const uint16_t *left, int bd";
    specialize qw/av1_highbd_dc_filter_predictor sse4_1/;
    add_proto qw/void av1_highbd_v_filter_predictor/, "uint16_t *dst, ptrdiff_t stride, int bs, const uint16_t *above, const uint16_t *left, int bd";
    specialize qw/av1_highbd_v_filter_predictor sse4_1/;
    add_proto qw/void av1_highbd_h_filter_predictor/, "uint16_t *dst, ptrdiff_t stride, int bs, const uint16_t *above, const uint16_t *left, int bd";
    specialize qw/av1_highbd_h_filter_predictor sse4_1/;
    add_proto qw/void av1_highbd_d45_filter_predictor/, "uint16_t *dst, ptrdiff_t stride, int bs, const uint16_t *above, const uint16_t *left, int bd";
    specialize qw/av1_highbd_d45_filter_predictor sse4_1/;
    add_proto qw/void av1_highbd_d135_filter_predictor/, "uint16_t *dst, ptrdiff_t stride, int bs, const uint16_t *above, const uint16_t *left, int bd";
    specialize qw/av1_highbd_d135_filter_predictor sse4_1/;
    add_proto qw/void av1_highbd_d117_filter_predictor/, "uint16_t *dst, ptrdiff_t stride, int bs, const uint16_t *above, const uint16_t *left, int bd";
    specialize qw/av1_highbd_d117_filter_predictor sse4_1/;
    add_proto qw/void av1_highbd_d153_filter_predictor/, "uint16_t *dst, ptrdiff_t stride, int bs, const uint16_t *above, const uint16_t *left, int bd";
    specialize qw/av1_highbd_d153_filter_predictor sse4_1/;
    add_proto qw/void av1_highbd_d207_filter_predictor/, "uint16_t *dst, ptrdiff_t stride, int bs, const uint16_t *above, const uint16_t *left, int bd";
    specialize qw/av1_highbd_d207_filter_predictor sse4_1/;
    add_proto qw/void av1_highbd_d63_filter_predictor/, "uint16_t *dst, ptrdiff_t stride, int bs, const uint16_t *above, const uint16_t *left, int bd";
    specialize qw/av1_highbd_d63_filter_predictor sse4_1/;
    add_proto qw/void av1_highbd_tm_filter_predictor/, "uint16_t *dst, ptrdiff_t stride, int bs, const uint16_t *above, const uint16_t *left, int bd";
    specialize qw/av1_highbd_tm_filter_predictor sse4_1/;
  }
}

# High bitdepth functions
if (aom_config("CONFIG_AOM_HIGHBITDEPTH") eq "yes") {
  #
  # Sub Pixel Filters
  #
  add_proto qw/void av1_highbd_convolve_copy/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
  specialize qw/av1_highbd_convolve_copy/;

  add_proto qw/void av1_highbd_convolve_avg/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
  specialize qw/av1_highbd_convolve_avg/;

  add_proto qw/void av1_highbd_convolve8/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
  specialize qw/av1_highbd_convolve8/, "$sse2_x86_64";

  add_proto qw/void av1_highbd_convolve8_horiz/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
  specialize qw/av1_highbd_convolve8_horiz/, "$sse2_x86_64";

  add_proto qw/void av1_highbd_convolve8_vert/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
  specialize qw/av1_highbd_convolve8_vert/, "$sse2_x86_64";

  add_proto qw/void av1_highbd_convolve8_avg/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
  specialize qw/av1_highbd_convolve8_avg/, "$sse2_x86_64";

  add_proto qw/void av1_highbd_convolve8_avg_horiz/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
  specialize qw/av1_highbd_convolve8_avg_horiz/, "$sse2_x86_64";

  add_proto qw/void av1_highbd_convolve8_avg_vert/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h, int bps";
  specialize qw/av1_highbd_convolve8_avg_vert/, "$sse2_x86_64";

  #
  # dct
  #
  # Note as optimized versions of these functions are added we need to add a check to ensure
  # that when CONFIG_EMULATE_HARDWARE is on, it defaults to the C versions only.
  add_proto qw/void av1_highbd_iht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type, int bd";
  specialize qw/av1_highbd_iht4x4_16_add/;

  add_proto qw/void av1_highbd_iht4x8_32_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type, int bd";
  specialize qw/av1_highbd_iht4x8_32_add/;

  add_proto qw/void av1_highbd_iht8x4_32_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type, int bd";
  specialize qw/av1_highbd_iht8x4_32_add/;

  add_proto qw/void av1_highbd_iht8x16_128_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type, int bd";
  specialize qw/av1_highbd_iht8x16_128_add/;

  add_proto qw/void av1_highbd_iht16x8_128_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type, int bd";
  specialize qw/av1_highbd_iht16x8_128_add/;

  add_proto qw/void av1_highbd_iht16x32_512_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type, int bd";
  specialize qw/av1_highbd_iht16x32_512_add/;

  add_proto qw/void av1_highbd_iht32x16_512_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type, int bd";
  specialize qw/av1_highbd_iht32x16_512_add/;

  add_proto qw/void av1_highbd_iht8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type, int bd";
  specialize qw/av1_highbd_iht8x8_64_add/;

  add_proto qw/void av1_highbd_iht16x16_256_add/, "const tran_low_t *input, uint8_t *output, int pitch, int tx_type, int bd";
  specialize qw/av1_highbd_iht16x16_256_add/;
}

#
# Encoder functions below this point.
#
if (aom_config("CONFIG_AV1_ENCODER") eq "yes") {

# ENCODEMB INVOKE

if (aom_config("CONFIG_AOM_QM") eq "yes") {
  if (aom_config("CONFIG_AOM_HIGHBITDEPTH") eq "yes") {
    # the transform coefficients are held in 32-bit
    # values, so the assembler code for  av1_block_error can no longer be used.
    add_proto qw/int64_t av1_block_error/, "const tran_low_t *coeff, const tran_low_t *dqcoeff, intptr_t block_size, int64_t *ssz";
    specialize qw/av1_block_error/;

    add_proto qw/void av1_quantize_fp/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan, const qm_val_t * qm_ptr, const qm_val_t *iqm_ptr";

    add_proto qw/void av1_quantize_fp_32x32/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan, const qm_val_t * qm_ptr, const qm_val_t *iqm_ptr";

    add_proto qw/void av1_fdct8x8_quant/, "const int16_t *input, int stride, tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan, const qm_val_t * qm_ptr, const qm_val_t *iqm_ptr";
    specialize qw/av1_fdct8x8_quant/;
  } else {
    add_proto qw/int64_t av1_block_error/, "const tran_low_t *coeff, const tran_low_t *dqcoeff, intptr_t block_size, int64_t *ssz";
    specialize qw/av1_block_error avx2 msa/, "$sse2_x86inc";

    add_proto qw/int64_t av1_block_error_fp/, "const int16_t *coeff, const int16_t *dqcoeff, int block_size";
    specialize qw/av1_block_error_fp neon/, "$sse2_x86inc";

    add_proto qw/void av1_quantize_fp/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan, const qm_val_t * qm_ptr, const qm_val_t *iqm_ptr";

    add_proto qw/void av1_quantize_fp_32x32/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan, const qm_val_t * qm_ptr, const qm_val_t *iqm_ptr";

    add_proto qw/void av1_fdct8x8_quant/, "const int16_t *input, int stride, tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan, const qm_val_t * qm_ptr, const qm_val_t *iqm_ptr";
  }
} else {
  if (aom_config("CONFIG_AOM_HIGHBITDEPTH") eq "yes") {
    # the transform coefficients are held in 32-bit
    # values, so the assembler code for  av1_block_error can no longer be used.
    add_proto qw/int64_t av1_block_error/, "const tran_low_t *coeff, const tran_low_t *dqcoeff, intptr_t block_size, int64_t *ssz";
    specialize qw/av1_block_error/;

    add_proto qw/void av1_quantize_fp/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan";
    specialize qw/av1_quantize_fp/;

    add_proto qw/void av1_quantize_fp_32x32/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan";
    specialize qw/av1_quantize_fp_32x32/;

    add_proto qw/void av1_fdct8x8_quant/, "const int16_t *input, int stride, tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan";
    specialize qw/av1_fdct8x8_quant/;
  } else {
    add_proto qw/int64_t av1_block_error/, "const tran_low_t *coeff, const tran_low_t *dqcoeff, intptr_t block_size, int64_t *ssz";
    specialize qw/av1_block_error sse2 avx2 msa/;

    add_proto qw/int64_t av1_block_error_fp/, "const int16_t *coeff, const int16_t *dqcoeff, int block_size";
    specialize qw/av1_block_error_fp neon sse2/;

    add_proto qw/void av1_quantize_fp/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan";
    specialize qw/av1_quantize_fp neon sse2/, "$ssse3_x86_64";

    add_proto qw/void av1_quantize_fp_32x32/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan";
    specialize qw/av1_quantize_fp_32x32/, "$ssse3_x86_64";

    add_proto qw/void av1_fdct8x8_quant/, "const int16_t *input, int stride, tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan";
    specialize qw/av1_fdct8x8_quant sse2 ssse3 neon/;
  }

}

# fdct functions

if (aom_config("CONFIG_AOM_HIGHBITDEPTH") eq "yes") {
  add_proto qw/void av1_fht4x4/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht4x4 sse2/;

  add_proto qw/void av1_fht4x8/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht4x8/;

  add_proto qw/void av1_fht8x4/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht8x4/;

  add_proto qw/void av1_fht8x16/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht8x16/;

  add_proto qw/void av1_fht16x8/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht16x8/;

  add_proto qw/void av1_fht16x32/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht16x32/;

  add_proto qw/void av1_fht32x16/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht32x16/;

  add_proto qw/void av1_fht8x8/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht8x8 sse2/;

  add_proto qw/void av1_fht16x16/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht16x16 sse2/;

  add_proto qw/void av1_fht32x32/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht32x32/;

  add_proto qw/void av1_fwht4x4/, "const int16_t *input, tran_low_t *output, int stride";
  specialize qw/av1_fwht4x4/;
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
  if (aom_config("CONFIG_EMULATE_HARDWARE") eq "yes") {
    add_proto qw/void av1_fdct4x4/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct4x4/;

    add_proto qw/void av1_fdct4x4_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct4x4_1/;

    add_proto qw/void av1_fdct8x8/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct8x8/;

    add_proto qw/void av1_fdct8x8_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct8x8_1/;

    add_proto qw/void av1_fdct16x16/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct16x16/;

    add_proto qw/void av1_fdct16x16_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct16x16_1/;

    add_proto qw/void av1_fdct32x32/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32/;

    add_proto qw/void av1_fdct32x32_rd/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32_rd/;

    add_proto qw/void av1_fdct32x32_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32_1/;

    add_proto qw/void av1_highbd_fdct4x4/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct4x4/;

    add_proto qw/void av1_highbd_fdct8x8/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct8x8/;

    add_proto qw/void av1_highbd_fdct8x8_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct8x8_1/;

    add_proto qw/void av1_highbd_fdct16x16/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct16x16/;

    add_proto qw/void av1_highbd_fdct16x16_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct16x16_1/;

    add_proto qw/void av1_highbd_fdct32x32/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct32x32/;

    add_proto qw/void av1_highbd_fdct32x32_rd/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct32x32_rd/;

    add_proto qw/void av1_highbd_fdct32x32_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct32x32_1/;
  } else {
    add_proto qw/void av1_fdct4x4/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct4x4 sse2/;

    add_proto qw/void av1_fdct4x4_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct4x4_1 sse2/;

    add_proto qw/void av1_fdct8x8/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct8x8 sse2/;

    add_proto qw/void av1_fdct8x8_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct8x8_1 sse2/;

    add_proto qw/void av1_fdct16x16/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct16x16 sse2/;

    add_proto qw/void av1_fdct16x16_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct16x16_1 sse2/;

    add_proto qw/void av1_fdct32x32/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32 sse2/;

    add_proto qw/void av1_fdct32x32_rd/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32_rd sse2/;

    add_proto qw/void av1_fdct32x32_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32_1 sse2/;

    add_proto qw/void av1_highbd_fdct4x4/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct4x4 sse2/;

    add_proto qw/void av1_highbd_fdct8x8/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct8x8 sse2/;

    add_proto qw/void av1_highbd_fdct8x8_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct8x8_1/;

    add_proto qw/void av1_highbd_fdct16x16/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct16x16 sse2/;

    add_proto qw/void av1_highbd_fdct16x16_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct16x16_1/;

    add_proto qw/void av1_highbd_fdct32x32/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct32x32 sse2/;

    add_proto qw/void av1_highbd_fdct32x32_rd/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct32x32_rd sse2/;

    add_proto qw/void av1_highbd_fdct32x32_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_highbd_fdct32x32_1/;
  }
Yaowu Xu's avatar
Yaowu Xu committed
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
} else {
  add_proto qw/void av1_fht4x4/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht4x4 sse2/;

  add_proto qw/void av1_fht4x8/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht4x8/;

  add_proto qw/void av1_fht8x4/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht8x4/;

  add_proto qw/void av1_fht8x16/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht8x16/;

  add_proto qw/void av1_fht16x8/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht16x8/;

  add_proto qw/void av1_fht16x32/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht16x32/;

  add_proto qw/void av1_fht32x16/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht32x16/;

  add_proto qw/void av1_fht8x8/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht8x8 sse2/;

  add_proto qw/void av1_fht16x16/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht16x16 sse2/;

  if (aom_config("CONFIG_EXT_TX") ne "yes") {
    specialize qw/av1_fht4x4 msa/;
    specialize qw/av1_fht8x8 msa/;
    specialize qw/av1_fht16x16 msa/;
  }

  add_proto qw/void av1_fht32x32/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_fht32x32/;

  add_proto qw/void av1_fwht4x4/, "const int16_t *input, tran_low_t *output, int stride";
  specialize qw/av1_fwht4x4/;
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
  if (aom_config("CONFIG_EMULATE_HARDWARE") eq "yes") {
    add_proto qw/void av1_fdct4x4/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct4x4/;

    add_proto qw/void av1_fdct4x4_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct4x4_1/;

    add_proto qw/void av1_fdct8x8/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct8x8/;

    add_proto qw/void av1_fdct8x8_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct8x8_1/;

    add_proto qw/void av1_fdct16x16/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct16x16/;

    add_proto qw/void av1_fdct16x16_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct16x16_1/;

    add_proto qw/void av1_fdct32x32/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32/;

    add_proto qw/void av1_fdct32x32_rd/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32_rd/;

    add_proto qw/void av1_fdct32x32_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32_1/;
  } else {
    add_proto qw/void av1_fdct4x4/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct4x4 sse2/;

    add_proto qw/void av1_fdct4x4_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct4x4_1 sse2/;

    add_proto qw/void av1_fdct8x8/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct8x8 sse2/;

    add_proto qw/void av1_fdct8x8_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct8x8_1 sse2/;

    add_proto qw/void av1_fdct16x16/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct16x16 sse2/;

    add_proto qw/void av1_fdct16x16_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct16x16_1 sse2/;

    add_proto qw/void av1_fdct32x32/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32 sse2/;

    add_proto qw/void av1_fdct32x32_rd/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32_rd sse2/;

    add_proto qw/void av1_fdct32x32_1/, "const int16_t *input, tran_low_t *output, int stride";
    specialize qw/av1_fdct32x32_1 sse2/;
  }
Yaowu Xu's avatar
Yaowu Xu committed
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
}

add_proto qw/void av1_fwd_idtx/, "const int16_t *src_diff, tran_low_t *coeff, int stride, int bs, int tx_type";
  specialize qw/av1_fwd_idtx/;

# Inverse transform
if (aom_config("CONFIG_AOM_HIGHBITDEPTH") eq "yes") {
  # Note as optimized versions of these functions are added we need to add a check to ensure
  # that when CONFIG_EMULATE_HARDWARE is on, it defaults to the C versions only.
  add_proto qw/void av1_idct4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct4x4_1_add/;

  add_proto qw/void av1_idct4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct4x4_16_add/;

  add_proto qw/void av1_idct8x8_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct8x8_1_add/;

  add_proto qw/void av1_idct8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct8x8_64_add/;

  add_proto qw/void av1_idct8x8_12_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct8x8_12_add/;

  add_proto qw/void av1_idct16x16_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct16x16_1_add/;

  add_proto qw/void av1_idct16x16_256_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct16x16_256_add/;

  add_proto qw/void av1_idct16x16_10_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct16x16_10_add/;

  add_proto qw/void av1_idct32x32_1024_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct32x32_1024_add/;

  add_proto qw/void av1_idct32x32_34_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct32x32_34_add/;

  add_proto qw/void av1_idct32x32_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_idct32x32_1_add/;

  add_proto qw/void av1_iwht4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_iwht4x4_1_add/;

  add_proto qw/void av1_iwht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
  specialize qw/av1_iwht4x4_16_add/;

  add_proto qw/void av1_highbd_idct4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
  specialize qw/av1_highbd_idct4x4_1_add/;

  add_proto qw/void av1_highbd_idct8x8_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
  specialize qw/av1_highbd_idct8x8_1_add/;

  add_proto qw/void av1_highbd_idct16x16_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
  specialize qw/av1_highbd_idct16x16_1_add/;

  add_proto qw/void av1_highbd_idct32x32_1024_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
  specialize qw/av1_highbd_idct32x32_1024_add/;

  add_proto qw/void av1_highbd_idct32x32_34_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
  specialize qw/av1_highbd_idct32x32_34_add/;

  add_proto qw/void av1_highbd_idct32x32_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
  specialize qw/av1_highbd_idct32x32_1_add/;

  add_proto qw/void av1_highbd_iwht4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
  specialize qw/av1_highbd_iwht4x4_1_add/;

  add_proto qw/void av1_highbd_iwht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
  specialize qw/av1_highbd_iwht4x4_16_add/;

  # Force C versions if CONFIG_EMULATE_HARDWARE is 1
  if (aom_config("CONFIG_EMULATE_HARDWARE") eq "yes") {
    add_proto qw/void av1_highbd_idct4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
    specialize qw/av1_highbd_idct4x4_16_add/;

    add_proto qw/void av1_highbd_idct8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
    specialize qw/av1_highbd_idct8x8_64_add/;

    add_proto qw/void av1_highbd_idct8x8_10_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
    specialize qw/av1_highbd_idct8x8_10_add/;

    add_proto qw/void av1_highbd_idct16x16_256_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
    specialize qw/av1_highbd_idct16x16_256_add/;

    add_proto qw/void av1_highbd_idct16x16_10_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
    specialize qw/av1_highbd_idct16x16_10_add/;
  } else {
    add_proto qw/void av1_highbd_idct4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
    specialize qw/av1_highbd_idct4x4_16_add sse2/;

    add_proto qw/void av1_highbd_idct8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
    specialize qw/av1_highbd_idct8x8_64_add sse2/;

    add_proto qw/void av1_highbd_idct8x8_10_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
    specialize qw/av1_highbd_idct8x8_10_add sse2/;

    add_proto qw/void av1_highbd_idct16x16_256_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
    specialize qw/av1_highbd_idct16x16_256_add sse2/;

    add_proto qw/void av1_highbd_idct16x16_10_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int bd";
    specialize qw/av1_highbd_idct16x16_10_add sse2/;
  }  # CONFIG_EMULATE_HARDWARE
} else {
  # Force C versions if CONFIG_EMULATE_HARDWARE is 1
  if (aom_config("CONFIG_EMULATE_HARDWARE") eq "yes") {
    add_proto qw/void av1_idct4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct4x4_1_add/;

    add_proto qw/void av1_idct4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct4x4_16_add/;

    add_proto qw/void av1_idct8x8_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct8x8_1_add/;

    add_proto qw/void av1_idct8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct8x8_64_add/;

    add_proto qw/void av1_idct8x8_12_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct8x8_12_add/;

    add_proto qw/void av1_idct16x16_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct16x16_1_add/;

    add_proto qw/void av1_idct16x16_256_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct16x16_256_add/;

    add_proto qw/void av1_idct16x16_10_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct16x16_10_add/;

    add_proto qw/void av1_idct32x32_1024_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct32x32_1024_add/;

    add_proto qw/void av1_idct32x32_34_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct32x32_34_add/;

    add_proto qw/void av1_idct32x32_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct32x32_1_add/;

    add_proto qw/void av1_iwht4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_iwht4x4_1_add/;

    add_proto qw/void av1_iwht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_iwht4x4_16_add/;
  } else {
    add_proto qw/void av1_idct4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct4x4_1_add sse2/;

    add_proto qw/void av1_idct4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct4x4_16_add sse2/;

    add_proto qw/void av1_idct8x8_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct8x8_1_add sse2/;

    add_proto qw/void av1_idct8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct8x8_64_add sse2/;

    add_proto qw/void av1_idct8x8_12_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct8x8_12_add sse2/;

    add_proto qw/void av1_idct16x16_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct16x16_1_add sse2/;

    add_proto qw/void av1_idct16x16_256_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct16x16_256_add sse2/;

    add_proto qw/void av1_idct16x16_10_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct16x16_10_add sse2/;

    add_proto qw/void av1_idct32x32_1024_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct32x32_1024_add sse2/;

    add_proto qw/void av1_idct32x32_34_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct32x32_34_add sse2/;

    add_proto qw/void av1_idct32x32_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_idct32x32_1_add sse2/;

    add_proto qw/void av1_iwht4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_iwht4x4_1_add/;

    add_proto qw/void av1_iwht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
    specialize qw/av1_iwht4x4_16_add/;
  }  # CONFIG_EMULATE_HARDWARE
}  # CONFIG_AOM_HIGHBITDEPTH

if (aom_config("CONFIG_AOM_HIGHBITDEPTH") eq "yes") {
  #fwd txfm
  add_proto qw/void av1_fwd_txfm2d_4x4/, "const int16_t *input, int32_t *output, int stride, int tx_type, int bd";
  specialize qw/av1_fwd_txfm2d_4x4 sse4_1/;
  add_proto qw/void av1_fwd_txfm2d_8x8/, "const int16_t *input, int32_t *output, int stride, int tx_type, int bd";
  specialize qw/av1_fwd_txfm2d_8x8 sse4_1/;
  add_proto qw/void av1_fwd_txfm2d_16x16/, "const int16_t *input, int32_t *output, int stride, int tx_type, int bd";
  specialize qw/av1_fwd_txfm2d_16x16 sse4_1/;
  add_proto qw/void av1_fwd_txfm2d_32x32/, "const int16_t *input, int32_t *output, int stride, int tx_type, int bd";
  specialize qw/av1_fwd_txfm2d_32x32 sse4_1/;
  add_proto qw/void av1_fwd_txfm2d_64x64/, "const int16_t *input, int32_t *output, int stride, int tx_type, int bd";
  specialize qw/av1_fwd_txfm2d_64x64 sse4_1/;

  #inv txfm
  add_proto qw/void av1_inv_txfm2d_add_4x4/, "const int32_t *input, uint16_t *output, int stride, int tx_type, int bd";
  specialize qw/av1_inv_txfm2d_add_4x4 sse4_1/;
  add_proto qw/void av1_inv_txfm2d_add_8x8/, "const int32_t *input, uint16_t *output, int stride, int tx_type, int bd";
  specialize qw/av1_inv_txfm2d_add_8x8 sse4_1/;
  add_proto qw/void av1_inv_txfm2d_add_16x16/, "const int32_t *input, uint16_t *output, int stride, int tx_type, int bd";
  specialize qw/av1_inv_txfm2d_add_16x16 sse4_1/;
  add_proto qw/void av1_inv_txfm2d_add_32x32/, "const int32_t *input, uint16_t *output, int stride, int tx_type, int bd";
  specialize qw/av1_inv_txfm2d_add_32x32/;
  add_proto qw/void av1_inv_txfm2d_add_64x64/, "const int32_t *input, uint16_t *output, int stride, int tx_type, int bd";
  specialize qw/av1_inv_txfm2d_add_64x64/;
}

#
# Motion search
#
add_proto qw/int av1_full_search_sad/, "const struct macroblock *x, const struct mv *ref_mv, int sad_per_bit, int distance, const struct aom_variance_vtable *fn_ptr, const struct mv *center_mv, struct mv *best_mv";
specialize qw/av1_full_search_sad sse3 sse4_1/;
$av1_full_search_sad_sse3=av1_full_search_sadx3;
$av1_full_search_sad_sse4_1=av1_full_search_sadx8;

add_proto qw/int av1_diamond_search_sad/, "struct macroblock *x, const struct search_site_config *cfg,  struct mv *ref_mv, struct mv *best_mv, int search_param, int sad_per_bit, int *num00, const struct aom_variance_vtable *fn_ptr, const struct mv *center_mv";
specialize qw/av1_diamond_search_sad/;

add_proto qw/int av1_full_range_search/, "const struct macroblock *x, const struct search_site_config *cfg, struct mv *ref_mv, struct mv *best_mv, int search_param, int sad_per_bit, int *num00, const struct aom_variance_vtable *fn_ptr, const struct mv *center_mv";
specialize qw/av1_full_range_search/;

add_proto qw/void av1_temporal_filter_apply/, "uint8_t *frame1, unsigned int stride, uint8_t *frame2, unsigned int block_width, unsigned int block_height, int strength, int filter_weight, unsigned int *accumulator, uint16_t *count";
specialize qw/av1_temporal_filter_apply sse2 msa/;

if (aom_config("CONFIG_AOM_HIGHBITDEPTH") eq "yes") {

  # ENCODEMB INVOKE
  if (aom_config("CONFIG_NEW_QUANT") eq "yes") {
    add_proto qw/void highbd_quantize_nuq/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, const int16_t *dequant_ptr, const cuml_bins_type_nuq *cuml_bins_ptr, const dequant_val_type_nuq *dequant_val, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, uint16_t *eob_ptr, const int16_t *scan, const uint8_t *band";
    specialize qw/highbd_quantize_nuq/;

    add_proto qw/void highbd_quantize_fp_nuq/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *quant_ptr, const int16_t *dequant_ptr, const cuml_bins_type_nuq *cuml_bins_ptr, const dequant_val_type_nuq *dequant_val, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, uint16_t *eob_ptr, const int16_t *scan, const uint8_t *band";
    specialize qw/highbd_quantize_fp_nuq/;

    add_proto qw/void highbd_quantize_32x32_nuq/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, const int16_t *dequant_ptr, const cuml_bins_type_nuq *cuml_bins_ptr, const dequant_val_type_nuq *dequant_val, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, uint16_t *eob_ptr, const int16_t *scan, const uint8_t *band";
    specialize qw/highbd_quantize_32x32_nuq/;

    add_proto qw/void highbd_quantize_32x32_fp_nuq/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *quant_ptr, const int16_t *dequant_ptr, const cuml_bins_type_nuq *cuml_bins_ptr, const dequant_val_type_nuq *dequant_val, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, uint16_t *eob_ptr, const int16_t *scan, const uint8_t *band";
    specialize qw/highbd_quantize_32x32_fp_nuq/;
  }

  add_proto qw/int64_t av1_highbd_block_error/, "const tran_low_t *coeff, const tran_low_t *dqcoeff, intptr_t block_size, int64_t *ssz, int bd";
  specialize qw/av1_highbd_block_error sse2/;

  if (aom_config("CONFIG_AOM_QM") eq "yes") {
    add_proto qw/void av1_highbd_quantize_fp/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan, const qm_val_t * qm_ptr, const qm_val_t * iqm_ptr";

    add_proto qw/void av1_highbd_quantize_fp_32x32/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan, const qm_val_t * qm_ptr, const qm_val_t * iqm_ptr";
  } else {
    add_proto qw/void av1_highbd_quantize_fp/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan, int log_scale";
    specialize qw/av1_highbd_quantize_fp sse4_1/;

    add_proto qw/void av1_highbd_quantize_b/, "const tran_low_t *coeff_ptr, intptr_t n_coeffs, int skip_block, const int16_t *zbin_ptr, const int16_t *round_ptr, const int16_t *quant_ptr, const int16_t *quant_shift_ptr, tran_low_t *qcoeff_ptr, tran_low_t *dqcoeff_ptr, const int16_t *dequant_ptr, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan, int log_scale";
    specialize qw/av1_highbd_quantize_b/;
  }

  # fdct functions
  add_proto qw/void av1_highbd_fht4x4/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_highbd_fht4x4 sse4_1/;

  add_proto qw/void av1_highbd_fht4x8/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_highbd_fht4x8/;

  add_proto qw/void av1_highbd_fht8x4/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_highbd_fht8x4/;

  add_proto qw/void av1_highbd_fht8x16/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_highbd_fht8x16/;

  add_proto qw/void av1_highbd_fht16x8/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_highbd_fht16x8/;

  add_proto qw/void av1_highbd_fht16x32/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_highbd_fht16x32/;

  add_proto qw/void av1_highbd_fht32x16/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_highbd_fht32x16/;

  add_proto qw/void av1_highbd_fht8x8/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_highbd_fht8x8/;

  add_proto qw/void av1_highbd_fht16x16/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_highbd_fht16x16/;

  add_proto qw/void av1_highbd_fht32x32/, "const int16_t *input, tran_low_t *output, int stride, int tx_type";
  specialize qw/av1_highbd_fht32x32/;

  add_proto qw/void av1_highbd_fwht4x4/, "const int16_t *input, tran_low_t *output, int stride";
  specialize qw/av1_highbd_fwht4x4/;

  add_proto qw/void av1_highbd_temporal_filter_apply/, "uint8_t *frame1, unsigned int stride, uint8_t *frame2, unsigned int block_width, unsigned int block_height, int strength, int filter_weight, unsigned int *accumulator, uint16_t *count";
  specialize qw/av1_highbd_temporal_filter_apply/;

}
# End av1_high encoder functions

if (aom_config("CONFIG_EXT_INTER") eq "yes") {
  add_proto qw/uint64_t av1_wedge_sse_from_residuals/, "const int16_t *r1, const int16_t *d, const uint8_t *m, int N";
  specialize qw/av1_wedge_sse_from_residuals sse2/;
  add_proto qw/int av1_wedge_sign_from_residuals/, "const int16_t *ds, const uint8_t *m, int N, int64_t limit";
  specialize qw/av1_wedge_sign_from_residuals sse2/;
  add_proto qw/void av1_wedge_compute_delta_squares/, "int16_t *d, const int16_t *a, const int16_t *b, int N";
  specialize qw/av1_wedge_compute_delta_squares sse2/;
}

}
# end encoder functions
1;