1 |
;/***************************************************************************** |
;/************************************************************************** |
2 |
; * |
; * |
3 |
; * XVID MPEG-4 VIDEO CODEC |
; * XVID MPEG-4 VIDEO CODEC |
4 |
; * xmm (extended mmx) sum of absolute difference |
; * xmm sum of absolute difference |
5 |
; * |
; * |
6 |
; * Copyright(C) 2002 Peter Ross <pross@xvid.org> |
; * This program is an implementation of a part of one or more MPEG-4 |
7 |
; * Copyright(C) 2002 Michael Militzer <michael@xvid.org> |
; * Video tools as specified in ISO/IEC 14496-2 standard. Those intending |
8 |
; * Copyright(C) 2002 Pascal Massimino <skal@planet-d.net> |
; * to use this software module in hardware or software products are |
9 |
|
; * advised that its use may infringe existing patents or copyrights, and |
10 |
|
; * any such use would be at such party's own risk. The original |
11 |
|
; * developer of this software module and his/her company, and subsequent |
12 |
|
; * editors and their companies, will have no liability for use of this |
13 |
|
; * software or modifications or derivatives thereof. |
14 |
; * |
; * |
15 |
; * This file is part of XviD, a free MPEG-4 video encoder/decoder |
; * This program is free software; you can redistribute it and/or modify |
16 |
; * |
; * it under the terms of the GNU General Public License as published by |
|
; * XviD is free software; you can redistribute it and/or modify it |
|
|
; * under the terms of the GNU General Public License as published by |
|
17 |
; * the Free Software Foundation; either version 2 of the License, or |
; * the Free Software Foundation; either version 2 of the License, or |
18 |
; * (at your option) any later version. |
; * (at your option) any later version. |
19 |
; * |
; * |
24 |
; * |
; * |
25 |
; * You should have received a copy of the GNU General Public License |
; * You should have received a copy of the GNU General Public License |
26 |
; * along with this program; if not, write to the Free Software |
; * along with this program; if not, write to the Free Software |
27 |
; * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
; * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
|
; * |
|
|
; * Under section 8 of the GNU General Public License, the copyright |
|
|
; * holders of XVID explicitly forbid distribution in the following |
|
|
; * countries: |
|
|
; * |
|
|
; * - Japan |
|
|
; * - United States of America |
|
28 |
; * |
; * |
29 |
; * Linking XviD statically or dynamically with other modules is making a |
; *************************************************************************/ |
30 |
; * combined work based on XviD. Thus, the terms and conditions of the |
|
31 |
; * GNU General Public License cover the whole combination. |
;/************************************************************************** |
|
; * |
|
|
; * As a special exception, the copyright holders of XviD give you |
|
|
; * permission to link XviD with independent modules that communicate with |
|
|
; * XviD solely through the VFW1.1 and DShow interfaces, regardless of the |
|
|
; * license terms of these independent modules, and to copy and distribute |
|
|
; * the resulting combined work under terms of your choice, provided that |
|
|
; * every copy of the combined work is accompanied by a complete copy of |
|
|
; * the source code of XviD (the version of XviD used to produce the |
|
|
; * combined work), being distributed under the terms of the GNU General |
|
|
; * Public License plus this exception. An independent module is a module |
|
|
; * which is not derived from or based on XviD. |
|
32 |
; * |
; * |
33 |
; * Note that people who make modified versions of XviD are not obligated |
; * History: |
|
; * to grant this special exception for their modified versions; it is |
|
|
; * their choice whether to do so. The GNU General Public License gives |
|
|
; * permission to release a modified version without this exception; this |
|
|
; * exception also makes it possible to release a modified version which |
|
|
; * carries forward this exception. |
|
34 |
; * |
; * |
35 |
; * $Id$ |
; * 23.07.2002 sad8bi_xmm; <pross@xvid.org> |
36 |
|
; * 04.06.2002 rewrote some funcs (XMM mainly) -Skal- |
37 |
|
; * 17.11.2001 bugfix and small improvement for dev16_xmm, |
38 |
|
; * removed terminate early in sad16_xmm (Isibaar) |
39 |
|
; * 12.11.2001 inital version; (c)2001 peter ross <pross@cs.rmit.edu.au> |
40 |
; * |
; * |
41 |
; *************************************************************************/ |
; *************************************************************************/ |
42 |
|
|
357 |
|
|
358 |
mov eax, [esp+ 4] ; Src |
mov eax, [esp+ 4] ; Src |
359 |
|
|
360 |
|
|
361 |
pxor mm5, mm5 ; sums |
pxor mm5, mm5 ; sums |
362 |
pxor mm6, mm6 |
pxor mm6, mm6 |
363 |
|
|
386 |
|
|
387 |
movd eax, mm6 |
movd eax, mm6 |
388 |
ret |
ret |
389 |
|
|
390 |
|
cglobal sad16v_xmm |
391 |
|
|
392 |
|
;=========================================================================== |
393 |
|
;int sad16v_xmm(const uint8_t * const cur, |
394 |
|
; const uint8_t * const ref, |
395 |
|
; const uint32_t stride, |
396 |
|
; int* sad8); |
397 |
|
;=========================================================================== |
398 |
|
align 16 |
399 |
|
sad16v_xmm: |
400 |
|
push ebx |
401 |
|
mov eax, [esp+4+ 4] ; Src1 |
402 |
|
mov edx, [esp+4+ 8] ; Src2 |
403 |
|
mov ecx, [esp+4+12] ; Stride |
404 |
|
mov ebx, [esp+4+16] ; sad ptr |
405 |
|
|
406 |
|
pxor mm5, mm5 ; accum1 |
407 |
|
pxor mm6, mm6 ; accum2 |
408 |
|
pxor mm7, mm7 ; total |
409 |
|
SAD_16x16_SSE |
410 |
|
SAD_16x16_SSE |
411 |
|
SAD_16x16_SSE |
412 |
|
SAD_16x16_SSE |
413 |
|
SAD_16x16_SSE |
414 |
|
SAD_16x16_SSE |
415 |
|
SAD_16x16_SSE |
416 |
|
SAD_16x16_SSE |
417 |
|
paddusw mm7, mm5 |
418 |
|
paddusw mm7, mm6 |
419 |
|
movd [ebx], mm5 |
420 |
|
movd [ebx+4], mm6 |
421 |
|
|
422 |
|
pxor mm5, mm5 ; accum1 |
423 |
|
pxor mm6, mm6 ; accum2 |
424 |
|
SAD_16x16_SSE |
425 |
|
SAD_16x16_SSE |
426 |
|
SAD_16x16_SSE |
427 |
|
SAD_16x16_SSE |
428 |
|
SAD_16x16_SSE |
429 |
|
SAD_16x16_SSE |
430 |
|
SAD_16x16_SSE |
431 |
|
SAD_16x16_SSE |
432 |
|
paddusw mm7, mm5 |
433 |
|
paddusw mm7, mm6 |
434 |
|
movd [ebx+8], mm5 |
435 |
|
movd [ebx+12], mm6 |
436 |
|
|
437 |
|
movd eax, mm7 |
438 |
|
pop ebx |
439 |
|
ret |
440 |
|
;-------- |
441 |
|
|
442 |
|
|