Parent Directory | Revision Log
Revision 651 -
(view)
(download)
Original Path: trunk/xvidcore/src/image/x86_asm/yv12_to_yuyv_mmx.asm
1 : | chl | 434 | ;/***************************************************************************** |
2 : | Isibaar | 3 | ; * |
3 : | chl | 434 | ; * XVID MPEG-4 VIDEO CODEC |
4 : | ; * mmx yuv planar to yuyv/uyvy conversion | ||
5 : | Isibaar | 3 | ; * |
6 : | chl | 434 | ; * Copyright(C) 2002 Peter Ross <pross@xvid.org> |
7 : | Isibaar | 3 | ; * |
8 : | edgomez | 651 | ; * This file is part of XviD, a free MPEG-4 video encoder/decoder |
9 : | Isibaar | 3 | ; * |
10 : | edgomez | 651 | ; * XviD is free software; you can redistribute it and/or modify it |
11 : | ; * under the terms of the GNU General Public License as published by | ||
12 : | chl | 434 | ; * the Free Software Foundation; either version 2 of the License, or |
13 : | ; * (at your option) any later version. | ||
14 : | Isibaar | 3 | ; * |
15 : | chl | 434 | ; * This program is distributed in the hope that it will be useful, |
16 : | ; * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
17 : | ; * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
18 : | ; * GNU General Public License for more details. | ||
19 : | Isibaar | 3 | ; * |
20 : | chl | 434 | ; * You should have received a copy of the GNU General Public License |
21 : | ; * along with this program; if not, write to the Free Software | ||
22 : | ; * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | ||
23 : | Isibaar | 3 | ; * |
24 : | edgomez | 651 | ; * Under section 8 of the GNU General Public License, the copyright |
25 : | ; * holders of XVID explicitly forbid distribution in the following | ||
26 : | ; * countries: | ||
27 : | ; * | ||
28 : | ; * - Japan | ||
29 : | ; * - United States of America | ||
30 : | ; * | ||
31 : | ; * Linking XviD statically or dynamically with other modules is making a | ||
32 : | ; * combined work based on XviD. Thus, the terms and conditions of the | ||
33 : | ; * GNU General Public License cover the whole combination. | ||
34 : | ; * | ||
35 : | ; * As a special exception, the copyright holders of XviD give you | ||
36 : | ; * permission to link XviD with independent modules that communicate with | ||
37 : | ; * XviD solely through the VFW1.1 and DShow interfaces, regardless of the | ||
38 : | ; * license terms of these independent modules, and to copy and distribute | ||
39 : | ; * the resulting combined work under terms of your choice, provided that | ||
40 : | ; * every copy of the combined work is accompanied by a complete copy of | ||
41 : | ; * the source code of XviD (the version of XviD used to produce the | ||
42 : | ; * combined work), being distributed under the terms of the GNU General | ||
43 : | ; * Public License plus this exception. An independent module is a module | ||
44 : | ; * which is not derived from or based on XviD. | ||
45 : | ; * | ||
46 : | ; * Note that people who make modified versions of XviD are not obligated | ||
47 : | ; * to grant this special exception for their modified versions; it is | ||
48 : | ; * their choice whether to do so. The GNU General Public License gives | ||
49 : | ; * permission to release a modified version without this exception; this | ||
50 : | ; * exception also makes it possible to release a modified version which | ||
51 : | ; * carries forward this exception. | ||
52 : | ; * | ||
53 : | ; * $Id: yv12_to_yuyv_mmx.asm,v 1.3 2002-11-17 00:20:30 edgomez Exp $ | ||
54 : | ; * | ||
55 : | chl | 434 | ; ****************************************************************************/ |
56 : | Isibaar | 3 | |
57 : | bits 32 | ||
58 : | |||
59 : | |||
60 : | section .data | ||
61 : | |||
62 : | %macro cglobal 1 | ||
63 : | %ifdef PREFIX | ||
64 : | global _%1 | ||
65 : | %define %1 _%1 | ||
66 : | %else | ||
67 : | global %1 | ||
68 : | %endif | ||
69 : | %endmacro | ||
70 : | |||
71 : | align 16 | ||
72 : | |||
73 : | |||
74 : | section .text | ||
75 : | |||
76 : | |||
77 : | ;=========================================================================== | ||
78 : | ; | ||
79 : | ; void yv12_to_uyvy_mmx( | ||
80 : | ; uint8_t * dst, | ||
81 : | ; int dst_stride, | ||
82 : | ; uint8_t * y_src, | ||
83 : | ; uint8_t * u_src, | ||
84 : | ; uint8_t * v_src, | ||
85 : | ; int y_stride, | ||
86 : | ; int uv_stride, | ||
87 : | ; int width, | ||
88 : | ; int height); | ||
89 : | ; | ||
90 : | ; width must be multiple of 8 | ||
91 : | ; ~10% faster than plain c | ||
92 : | ; | ||
93 : | ;=========================================================================== | ||
94 : | |||
95 : | align 16 | ||
96 : | cglobal yv12_to_yuyv_mmx | ||
97 : | yv12_to_yuyv_mmx | ||
98 : | |||
99 : | push ebx | ||
100 : | push ecx | ||
101 : | push esi | ||
102 : | push edi | ||
103 : | push ebp ; STACK BASE = 20 | ||
104 : | |||
105 : | ; global constants | ||
106 : | |||
107 : | mov ebx, [esp + 20 + 32] ; width | ||
108 : | mov eax, [esp + 20 + 8] ; dst_stride | ||
109 : | sub eax, ebx ; | ||
110 : | add eax, eax ; eax = 2*(dst_stride - width) | ||
111 : | push eax ; [esp + 4] = dst_dif | ||
112 : | ; STACK BASE = 24 | ||
113 : | |||
114 : | shr ebx, 3 ; ebx = width / 8 | ||
115 : | mov edi, [esp + 24 + 4] ; dst | ||
116 : | |||
117 : | |||
118 : | ; --------- flip ------------- | ||
119 : | |||
120 : | mov ebp, [esp + 24 + 36] | ||
121 : | test ebp, ebp | ||
122 : | jl .flip | ||
123 : | |||
124 : | mov esi, [esp + 24 + 12] ; y_src | ||
125 : | mov ecx, [esp + 24 + 16] ; u_src | ||
126 : | mov edx, [esp + 24 + 20] ; v_src | ||
127 : | shr ebp, 1 ; y = height / 2 | ||
128 : | jmp short .yloop | ||
129 : | |||
130 : | |||
131 : | .flip | ||
132 : | neg ebp ; height = -height | ||
133 : | |||
134 : | mov eax, [esp + 24 + 24] ; y_stride | ||
135 : | lea edx, [ebp - 1] ; edx = height - 1 | ||
136 : | mul edx | ||
137 : | mov esi, [esp + 24 + 12] ; y_src | ||
138 : | add esi, eax ; y_src += (height - 1) * y_stride | ||
139 : | |||
140 : | shr ebp, 1 ; y = height / 2 | ||
141 : | mov eax, [esp + 24 + 28] ; uv_stride | ||
142 : | lea edx, [ebp - 1] ; edx = height/2 - 1 | ||
143 : | mul edx | ||
144 : | |||
145 : | mov ecx, [esp + 24 + 16] ; u_src | ||
146 : | mov edx, [esp + 24 + 20] ; v_src | ||
147 : | add ecx, eax ; u_src += (height/2 - 1) * uv_stride | ||
148 : | add edx, eax ; v_src += (height/2 - 1) * uv_stride | ||
149 : | |||
150 : | neg dword [esp + 24 + 24] ; y_stride = -y_stride | ||
151 : | neg dword [esp + 24 + 28] ; uv_stride = -uv_stride | ||
152 : | |||
153 : | .yloop | ||
154 : | xor eax, eax ; x = 0; | ||
155 : | |||
156 : | .xloop1 | ||
157 : | movd mm0, [ecx+4*eax] ; [ |uuuu] | ||
158 : | movd mm1, [edx+4*eax] ; [ |vvvv] | ||
159 : | movq mm2, [esi+8*eax] ; [yyyy|yyyy] | ||
160 : | |||
161 : | punpcklbw mm0, mm1 ; [vuvu|vuvu] | ||
162 : | movq mm3, mm2 | ||
163 : | punpcklbw mm2, mm0 ; [vyuy|vyuy] | ||
164 : | punpckhbw mm3, mm0 ; [vyuy|vyuy] | ||
165 : | movq [edi], mm2 | ||
166 : | movq [edi+8], mm3 | ||
167 : | |||
168 : | inc eax | ||
169 : | add edi, 16 | ||
170 : | |||
171 : | cmp eax, ebx | ||
172 : | jb .xloop1 | ||
173 : | |||
174 : | add edi, [esp + 0] ; dst += dst_dif | ||
175 : | add esi, [esp + 24 + 24] ; y_src += y_stride | ||
176 : | |||
177 : | xor eax, eax | ||
178 : | |||
179 : | .xloop2 | ||
180 : | movd mm0, [ecx+4*eax] ; [ |uuuu] | ||
181 : | movd mm1, [edx+4*eax] ; [ |vvvv] | ||
182 : | movq mm2, [esi+8*eax] ; [yyyy|yyyy] | ||
183 : | |||
184 : | punpcklbw mm0, mm1 ; [vuvu|vuvu] | ||
185 : | movq mm3, mm2 | ||
186 : | punpcklbw mm2, mm0 ; [vyuy|vyuy] | ||
187 : | punpckhbw mm3, mm0 ; [vyuy|vyuy] | ||
188 : | movq [edi], mm2 | ||
189 : | movq [edi+8], mm3 | ||
190 : | |||
191 : | inc eax | ||
192 : | add edi, 16 | ||
193 : | |||
194 : | cmp eax, ebx | ||
195 : | jb .xloop2 | ||
196 : | |||
197 : | add edi, [esp + 0] ; dst += dst_dif | ||
198 : | add esi, [esp + 24 + 24] ; y_src += y_stride | ||
199 : | add ecx, [esp + 24 + 28] ; u_src += uv_stride | ||
200 : | add edx, [esp + 24 + 28] ; v_src += uv_stride | ||
201 : | |||
202 : | dec ebp ; y-- | ||
203 : | jnz near .yloop | ||
204 : | |||
205 : | emms | ||
206 : | |||
207 : | add esp, 4 | ||
208 : | pop ebp | ||
209 : | pop edi | ||
210 : | pop esi | ||
211 : | pop ecx | ||
212 : | pop ebx | ||
213 : | |||
214 : | ret | ||
215 : | |||
216 : | |||
217 : | |||
218 : | |||
219 : | |||
220 : | ;=========================================================================== | ||
221 : | ; | ||
222 : | ; void yv12_to_uyvy_mmx( | ||
223 : | ; uint8_t * dst, | ||
224 : | ; int dst_stride, | ||
225 : | ; uint8_t * y_src, | ||
226 : | ; uint8_t * u_src, | ||
227 : | ; uint8_t * v_src, | ||
228 : | ; int y_stride, | ||
229 : | ; int uv_stride, | ||
230 : | ; int width, | ||
231 : | ; int height); | ||
232 : | ; | ||
233 : | ; width must be multiple of 8 | ||
234 : | ; ~10% faster than plain c | ||
235 : | ; | ||
236 : | ;=========================================================================== | ||
237 : | |||
238 : | align 16 | ||
239 : | cglobal yv12_to_uyvy_mmx | ||
240 : | yv12_to_uyvy_mmx | ||
241 : | |||
242 : | push ebx | ||
243 : | push ecx | ||
244 : | push esi | ||
245 : | push edi | ||
246 : | push ebp ; STACK BASE = 20 | ||
247 : | |||
248 : | ; global constants | ||
249 : | |||
250 : | mov ebx, [esp + 20 + 32] ; width | ||
251 : | mov eax, [esp + 20 + 8] ; dst_stride | ||
252 : | sub eax, ebx ; | ||
253 : | add eax, eax ; eax = 2*(dst_stride - width) | ||
254 : | push eax ; [esp + 4] = dst_dif | ||
255 : | ; STACK BASE = 24 | ||
256 : | |||
257 : | shr ebx, 3 ; ebx = width / 8 | ||
258 : | mov edi, [esp + 24 + 4] ; dst | ||
259 : | |||
260 : | |||
261 : | ; --------- flip ------------- | ||
262 : | |||
263 : | mov ebp, [esp + 24 + 36] | ||
264 : | test ebp, ebp | ||
265 : | jl .flip | ||
266 : | |||
267 : | mov esi, [esp + 24 + 12] ; y_src | ||
268 : | mov ecx, [esp + 24 + 16] ; u_src | ||
269 : | mov edx, [esp + 24 + 20] ; v_src | ||
270 : | shr ebp, 1 ; y = height / 2 | ||
271 : | jmp short .yloop | ||
272 : | |||
273 : | |||
274 : | .flip | ||
275 : | neg ebp ; height = -height | ||
276 : | |||
277 : | mov eax, [esp + 24 + 24] ; y_stride | ||
278 : | lea edx, [ebp - 1] ; edx = height - 1 | ||
279 : | mul edx | ||
280 : | mov esi, [esp + 24 + 12] ; y_src | ||
281 : | add esi, eax ; y_src += (height - 1) * y_stride | ||
282 : | |||
283 : | shr ebp, 1 ; y = height / 2 | ||
284 : | mov eax, [esp + 24 + 28] ; uv_stride | ||
285 : | lea edx, [ebp - 1] ; edx = height/2 - 1 | ||
286 : | mul edx | ||
287 : | |||
288 : | mov ecx, [esp + 24 + 16] ; u_src | ||
289 : | mov edx, [esp + 24 + 20] ; v_src | ||
290 : | add ecx, eax ; u_src += (height/2 - 1) * uv_stride | ||
291 : | add edx, eax ; v_src += (height/2 - 1) * uv_stride | ||
292 : | |||
293 : | neg dword [esp + 24 + 24] ; y_stride = -y_stride | ||
294 : | neg dword [esp + 24 + 28] ; uv_stride = -uv_stride | ||
295 : | |||
296 : | .yloop | ||
297 : | xor eax, eax ; x = 0; | ||
298 : | |||
299 : | .xloop1 | ||
300 : | movd mm0, [ecx+4*eax] ; [ |uuuu] | ||
301 : | movd mm1, [edx+4*eax] ; [ |vvvv] | ||
302 : | movq mm2, [esi+8*eax] ; [yyyy|yyyy] | ||
303 : | |||
304 : | punpcklbw mm0, mm1 ; [vuvu|vuvu] | ||
305 : | movq mm1, mm0 | ||
306 : | punpcklbw mm0, mm2 ; [yvyu|yvyu] | ||
307 : | punpckhbw mm1, mm2 ; [yvyu|yvyu] | ||
308 : | movq [edi], mm0 | ||
309 : | movq [edi+8], mm1 | ||
310 : | |||
311 : | inc eax | ||
312 : | add edi, 16 | ||
313 : | |||
314 : | cmp eax, ebx | ||
315 : | jb .xloop1 | ||
316 : | |||
317 : | add edi, [esp + 0] ; dst += dst_dif | ||
318 : | add esi, [esp + 24 + 24] ; y_src += y_stride | ||
319 : | |||
320 : | xor eax, eax | ||
321 : | |||
322 : | .xloop2 | ||
323 : | movd mm0, [ecx+4*eax] ; [ |uuuu] | ||
324 : | movd mm1, [edx+4*eax] ; [ |vvvv] | ||
325 : | movq mm2, [esi+8*eax] ; [yyyy|yyyy] | ||
326 : | |||
327 : | punpcklbw mm0, mm1 ; [vuvu|vuvu] | ||
328 : | movq mm1, mm0 | ||
329 : | punpcklbw mm0, mm2 ; [yvyu|yvyu] | ||
330 : | punpckhbw mm1, mm2 ; [yvyu|yvyu] | ||
331 : | |||
332 : | movq [edi], mm0 | ||
333 : | movq [edi+8], mm1 | ||
334 : | |||
335 : | inc eax | ||
336 : | add edi, 16 | ||
337 : | |||
338 : | cmp eax, ebx | ||
339 : | jb .xloop2 | ||
340 : | |||
341 : | add edi, [esp + 0] ; dst += dst_dif | ||
342 : | add esi, [esp + 24 + 24] ; y_src += y_stride | ||
343 : | add ecx, [esp + 24 + 28] ; u_src += uv_stride | ||
344 : | add edx, [esp + 24 + 28] ; v_src += uv_stride | ||
345 : | |||
346 : | dec ebp ; y-- | ||
347 : | jnz near .yloop | ||
348 : | |||
349 : | emms | ||
350 : | |||
351 : | add esp, 4 | ||
352 : | pop ebp | ||
353 : | pop edi | ||
354 : | pop esi | ||
355 : | pop ecx | ||
356 : | pop ebx | ||
357 : | |||
358 : | ret |
No admin address has been configured | ViewVC Help |
Powered by ViewVC 1.0.4 |