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
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
|
;******************************************************************************
;* H.264 intra prediction asm optimizations
;* Copyright (c) 2010 Jason Garrett-Glaser
;*
;* 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
;* 51, Inc., Foundation Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
;******************************************************************************
%include "x86inc.asm"
SECTION_RODATA
tm_shuf: times 8 db 0x03, 0x80
SECTION .text
cextern pb_3
;-----------------------------------------------------------------------------
; void pred16x16_vertical(uint8_t *src, int stride)
;-----------------------------------------------------------------------------
cglobal pred16x16_vertical_mmx, 2,3
sub r0, r1
mov r2, 8
movq mm0, [r0+0]
movq mm1, [r0+8]
.loop:
movq [r0+r1*1+0], mm0
movq [r0+r1*1+8], mm1
movq [r0+r1*2+0], mm0
movq [r0+r1*2+8], mm1
lea r0, [r0+r1*2]
dec r2
jg .loop
REP_RET
cglobal pred16x16_vertical_sse, 2,3
sub r0, r1
mov r2, 4
movaps xmm0, [r0]
.loop:
movaps [r0+r1*1], xmm0
movaps [r0+r1*2], xmm0
lea r0, [r0+r1*2]
movaps [r0+r1*1], xmm0
movaps [r0+r1*2], xmm0
lea r0, [r0+r1*2]
dec r2
jg .loop
REP_RET
;-----------------------------------------------------------------------------
; void pred16x16_horizontal(uint8_t *src, int stride)
;-----------------------------------------------------------------------------
%macro PRED16x16_H 1
cglobal pred16x16_horizontal_%1, 2,3
mov r2, 8
%ifidn %1, ssse3
mova m2, [pb_3]
%endif
.loop:
movd m0, [r0+r1*0-4]
movd m1, [r0+r1*1-4]
%ifidn %1, ssse3
pshufb m0, m2
pshufb m1, m2
%else
punpcklbw m0, m0
punpcklbw m1, m1
%ifidn %1, mmxext
pshufw m0, m0, 0xff
pshufw m1, m1, 0xff
%else
punpckhwd m0, m0
punpckhwd m1, m1
punpckhdq m0, m0
punpckhdq m1, m1
%endif
mova [r0+r1*0+8], m0
mova [r0+r1*1+8], m1
%endif
mova [r0+r1*0], m0
mova [r0+r1*1], m1
lea r0, [r0+r1*2]
dec r2
jg .loop
REP_RET
%endmacro
INIT_MMX
PRED16x16_H mmx
PRED16x16_H mmxext
INIT_XMM
PRED16x16_H ssse3
;-----------------------------------------------------------------------------
; void pred16x16_dc(uint8_t *src, int stride)
;-----------------------------------------------------------------------------
%macro PRED16x16_DC 2
cglobal pred16x16_dc_%1, 2,7
mov r4, r0
sub r0, r1
pxor mm0, mm0
pxor mm1, mm1
psadbw mm0, [r0+0]
psadbw mm1, [r0+8]
dec r0
movzx r5d, byte [r0+r1*1]
paddw mm0, mm1
movd r6d, mm0
lea r0, [r0+r1*2]
%rep 7
movzx r2d, byte [r0+r1*0]
movzx r3d, byte [r0+r1*1]
add r5d, r2d
add r6d, r3d
lea r0, [r0+r1*2]
%endrep
movzx r2d, byte [r0+r1*0]
add r5d, r6d
lea r2d, [r2+r5+16]
shr r2d, 5
%ifidn %1, mmx
movd m0, r2d
punpcklbw m0, m0
punpcklwd m0, m0
punpckldq m0, m0
%elifidn %1, mmxext
movd m0, r2d
punpcklbw m0, m0
pshufw m0, m0, 0
%elifidn %1, sse
imul r2d, 0x01010101
movd m0, r2d
shufps m0, m0, 0
%elifidn %1, sse2
movd m0, r2d
punpcklbw m0, m0
pshuflw m0, m0, 0
punpcklqdq m0, m0
%elifidn %1, ssse3
pxor m1, m1
movd m0, r2d
pshufb m0, m1
%endif
%if mmsize==8
mov r3d, 8
.loop:
%2 [r4+r1*0+0], m0
%2 [r4+r1*0+8], m0
%2 [r4+r1*1+0], m0
%2 [r4+r1*1+8], m0
%else
mov r3d, 4
.loop:
%2 [r4+r1*0], m0
%2 [r4+r1*1], m0
lea r4, [r4+r1*2]
%2 [r4+r1*0], m0
%2 [r4+r1*1], m0
%endif
lea r4, [r4+r1*2]
dec r3d
jg .loop
REP_RET
%endmacro
INIT_MMX
PRED16x16_DC mmx, movq
PRED16x16_DC mmxext, movq
INIT_XMM
PRED16x16_DC sse, movaps
PRED16x16_DC sse2, movdqa
PRED16x16_DC ssse3, movdqa
;-----------------------------------------------------------------------------
; void pred16x16_tm_vp8(uint8_t *src, int stride)
;-----------------------------------------------------------------------------
%macro PRED16x16_TM_MMX 1
cglobal pred16x16_tm_vp8_%1, 2,5
sub r0, r1
pxor mm7, mm7
movq mm0, [r0+0]
movq mm2, [r0+8]
movq mm1, mm0
movq mm3, mm2
punpcklbw mm0, mm7
punpckhbw mm1, mm7
punpcklbw mm2, mm7
punpckhbw mm3, mm7
movzx r3d, byte [r0-1]
mov r4d, 16
.loop:
movzx r2d, byte [r0+r1-1]
sub r2d, r3d
movd mm4, r2d
%ifidn %1, mmx
punpcklwd mm4, mm4
punpckldq mm4, mm4
%else
pshufw mm4, mm4, 0
%endif
movq mm5, mm4
movq mm6, mm4
movq mm7, mm4
paddw mm4, mm0
paddw mm5, mm1
paddw mm6, mm2
paddw mm7, mm3
packuswb mm4, mm5
packuswb mm6, mm7
movq [r0+r1+0], mm4
movq [r0+r1+8], mm6
add r0, r1
dec r4d
jg .loop
REP_RET
%endmacro
PRED16x16_TM_MMX mmx
PRED16x16_TM_MMX mmxext
cglobal pred16x16_tm_vp8_sse2, 2,6,6
sub r0, r1
pxor xmm2, xmm2
movdqa xmm0, [r0]
movdqa xmm1, xmm0
punpcklbw xmm0, xmm2
punpckhbw xmm1, xmm2
movzx r4d, byte [r0-1]
mov r5d, 8
.loop:
movzx r2d, byte [r0+r1*1-1]
movzx r3d, byte [r0+r1*2-1]
sub r2d, r4d
sub r3d, r4d
movd xmm2, r2d
movd xmm4, r3d
pshuflw xmm2, xmm2, 0
pshuflw xmm4, xmm4, 0
punpcklqdq xmm2, xmm2
punpcklqdq xmm4, xmm4
movdqa xmm3, xmm2
movdqa xmm5, xmm4
paddw xmm2, xmm0
paddw xmm3, xmm1
paddw xmm4, xmm0
paddw xmm5, xmm1
packuswb xmm2, xmm3
packuswb xmm4, xmm5
movdqa [r0+r1*1], xmm2
movdqa [r0+r1*2], xmm4
lea r0, [r0+r1*2]
dec r5d
jg .loop
REP_RET
;-----------------------------------------------------------------------------
; void pred8x8_vertical(uint8_t *src, int stride)
;-----------------------------------------------------------------------------
cglobal pred8x8_vertical_mmx, 2,2
sub r0, r1
movq mm0, [r0]
%rep 3
movq [r0+r1*1], mm0
movq [r0+r1*2], mm0
lea r0, [r0+r1*2]
%endrep
movq [r0+r1*1], mm0
movq [r0+r1*2], mm0
RET
;-----------------------------------------------------------------------------
; void pred8x8_horizontal(uint8_t *src, int stride)
;-----------------------------------------------------------------------------
%macro PRED8x8_H 1
cglobal pred8x8_horizontal_%1, 2,3
mov r2, 4
%ifidn %1, ssse3
mova m2, [pb_3]
%endif
.loop:
movd m0, [r0+r1*0-4]
movd m1, [r0+r1*1-4]
%ifidn %1, ssse3
pshufb m0, m2
pshufb m1, m2
%else
punpcklbw m0, m0
punpcklbw m1, m1
%ifidn %1, mmxext
pshufw m0, m0, 0xff
pshufw m1, m1, 0xff
%else
punpckhwd m0, m0
punpckhwd m1, m1
punpckhdq m0, m0
punpckhdq m1, m1
%endif
%endif
mova [r0+r1*0], m0
mova [r0+r1*1], m1
lea r0, [r0+r1*2]
dec r2
jg .loop
REP_RET
%endmacro
INIT_MMX
PRED8x8_H mmx
PRED8x8_H mmxext
PRED8x8_H ssse3
;-----------------------------------------------------------------------------
; void pred8x8_dc_rv40(uint8_t *src, int stride)
;-----------------------------------------------------------------------------
%macro PRED8x8_DC 1
cglobal pred8x8_dc_rv40_%1, 2,7
mov r4, r0
sub r0, r1
pxor mm0, mm0
psadbw mm0, [r0]
dec r0
movzx r5d, byte [r0+r1*1]
movd r6d, mm0
lea r0, [r0+r1*2]
%rep 3
movzx r2d, byte [r0+r1*0]
movzx r3d, byte [r0+r1*1]
add r5d, r2d
add r6d, r3d
lea r0, [r0+r1*2]
%endrep
movzx r2d, byte [r0+r1*0]
add r5d, r6d
lea r2d, [r2+r5+8]
shr r2d, 4
%ifidn %1, mmx
movd mm0, r2d
punpcklbw mm0, mm0
punpcklwd mm0, mm0
punpckldq mm0, mm0
%else
movd mm0, r2d
punpcklbw mm0, mm0
pshufw mm0, mm0, 0
%endif
mov r3d, 4
.loop:
movq [r4+r1*0], mm0
movq [r4+r1*1], mm0
lea r4, [r4+r1*2]
dec r3d
jg .loop
REP_RET
%endmacro
PRED8x8_DC mmx
PRED8x8_DC mmxext
;-----------------------------------------------------------------------------
; void pred8x8_tm_vp8(uint8_t *src, int stride)
;-----------------------------------------------------------------------------
%macro PRED8x8_TM_MMX 1
cglobal pred8x8_tm_vp8_%1, 2,6
sub r0, r1
pxor mm7, mm7
movq mm0, [r0]
movq mm1, mm0
punpcklbw mm0, mm7
punpckhbw mm1, mm7
movzx r4d, byte [r0-1]
mov r5d, 4
.loop:
movzx r2d, byte [r0+r1*1-1]
movzx r3d, byte [r0+r1*2-1]
sub r2d, r4d
sub r3d, r4d
movd mm2, r2d
movd mm4, r3d
%ifidn %1, mmx
punpcklwd mm2, mm2
punpcklwd mm4, mm4
punpckldq mm2, mm2
punpckldq mm4, mm4
%else
pshufw mm2, mm2, 0
pshufw mm4, mm4, 0
%endif
movq mm3, mm2
movq mm5, mm4
paddw mm2, mm0
paddw mm3, mm1
paddw mm4, mm0
paddw mm5, mm1
packuswb mm2, mm3
packuswb mm4, mm5
movq [r0+r1*1], mm2
movq [r0+r1*2], mm4
lea r0, [r0+r1*2]
dec r5d
jg .loop
REP_RET
%endmacro
PRED8x8_TM_MMX mmx
PRED8x8_TM_MMX mmxext
cglobal pred8x8_tm_vp8_sse2, 2,6,4
sub r0, r1
pxor xmm1, xmm1
movq xmm0, [r0]
punpcklbw xmm0, xmm1
movzx r4d, byte [r0-1]
mov r5d, 4
.loop:
movzx r2d, byte [r0+r1*1-1]
movzx r3d, byte [r0+r1*2-1]
sub r2d, r4d
sub r3d, r4d
movd xmm2, r2d
movd xmm3, r3d
pshuflw xmm2, xmm2, 0
pshuflw xmm3, xmm3, 0
punpcklqdq xmm2, xmm2
punpcklqdq xmm3, xmm3
paddw xmm2, xmm0
paddw xmm3, xmm0
packuswb xmm2, xmm3
movq [r0+r1*1], xmm2
movhps [r0+r1*2], xmm2
lea r0, [r0+r1*2]
dec r5d
jg .loop
REP_RET
cglobal pred8x8_tm_vp8_ssse3, 2,3,6
sub r0, r1
movdqa xmm4, [tm_shuf]
pxor xmm1, xmm1
movq xmm0, [r0]
punpcklbw xmm0, xmm1
movd xmm5, [r0-4]
pshufb xmm5, xmm4
mov r2d, 4
.loop:
movd xmm2, [r0+r1*1-4]
movd xmm3, [r0+r1*2-4]
pshufb xmm2, xmm4
pshufb xmm3, xmm4
psubw xmm2, xmm5
psubw xmm3, xmm5
paddw xmm2, xmm0
paddw xmm3, xmm0
packuswb xmm2, xmm3
movq [r0+r1*1], xmm2
movhps [r0+r1*2], xmm2
lea r0, [r0+r1*2]
dec r2d
jg .loop
REP_RET
|
