Parent Directory
| 
Revision Log
Revision 651 - (view) (download)
| 1 : | chl | 434 | ;/***************************************************************************** |
| 2 : | ; * | ||
| 3 : | ; * XVID MPEG-4 VIDEO CODEC | ||
| 4 : | ; * mmx yuv planar to yv12 conversion | ||
| 5 : | ; * | ||
| 6 : | ; * Copyright (C) 2001 - Michael Militzer <isibaar@xvid.org> | ||
| 7 : | ; * | ||
| 8 : | edgomez | 651 | ; * This file is part of XviD, a free MPEG-4 video encoder/decoder |
| 9 : | chl | 434 | ; * |
| 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 : | ; * | ||
| 15 : | ; * 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 : | ; * | ||
| 20 : | ; * 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 : | ; * | ||
| 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: yuv_to_yv12_mmx.asm,v 1.7 2002-11-17 00:20:30 edgomez Exp $ | ||
| 54 : | ; * | ||
| 55 : | chl | 434 | ; ****************************************************************************/ |
| 56 : | edgomez | 328 | |
| 57 : | BITS 32 | ||
| 58 : | |||
| 59 : | %macro cglobal 1 | ||
| 60 : | %ifdef PREFIX | ||
| 61 : | global _%1 | ||
| 62 : | edgomez | 331 | %define %1 _%1 |
| 63 : | edgomez | 328 | %else |
| 64 : | global %1 | ||
| 65 : | %endif | ||
| 66 : | %endmacro | ||
| 67 : | |||
| 68 : | SECTION .text | ||
| 69 : | |||
| 70 : | ALIGN 64 | ||
| 71 : | |||
| 72 : | ;------------------------------------------------------------------------------ | ||
| 73 : | ; | ||
| 74 : | ; void yuv_to_yv12_xmm(uint8_t *y_out, | ||
| 75 : | ; uint8_t *u_out, | ||
| 76 : | ; uint8_t *v_out, | ||
| 77 : | ; uint8_t *src, | ||
| 78 : | ; int width, int height, int stride); | ||
| 79 : | ; | ||
| 80 : | ; This function probably also runs on PentiumII class cpu's | ||
| 81 : | ; | ||
| 82 : | ; Attention: This code assumes that width is a multiple of 16 | ||
| 83 : | ; | ||
| 84 : | ;------------------------------------------------------------------------------ | ||
| 85 : | |||
| 86 : | |||
| 87 : | cglobal yuv_to_yv12_xmm | ||
| 88 : | yuv_to_yv12_xmm: | ||
| 89 : | |||
| 90 : | push ebx | ||
| 91 : | push esi | ||
| 92 : | push edi | ||
| 93 : | push ebp | ||
| 94 : | |||
| 95 : | ; local vars allocation | ||
| 96 : | %define localsize 4 | ||
| 97 : | %define remainder esp | ||
| 98 : | sub esp, localsize | ||
| 99 : | |||
| 100 : | ; function code | ||
| 101 : | mov eax, [esp + 40 + localsize] ; height -> eax | ||
| 102 : | mov ebx, [esp + 44 + localsize] ; stride -> ebx | ||
| 103 : | mov esi, [esp + 32 + localsize] ; src -> esi | ||
| 104 : | mov edi, [esp + 20 + localsize] ; y_out -> edi | ||
| 105 : | mov ecx, [esp + 36 + localsize] ; width -> ecx | ||
| 106 : | edgomez | 331 | |
| 107 : | edgomez | 328 | sub ebx, ecx ; stride - width -> ebx |
| 108 : | |||
| 109 : | mov edx, ecx | ||
| 110 : | mov ebp, ecx | ||
| 111 : | shr edx, 6 | ||
| 112 : | mov ecx, edx ; 64 bytes copied per iteration | ||
| 113 : | shl edx, 6 | ||
| 114 : | sub ebp, edx ; remainder -> ebp | ||
| 115 : | shr ebp, 4 ; 16 bytes per iteration | ||
| 116 : | add ebp, 1 | ||
| 117 : | mov [remainder], ebp | ||
| 118 : | |||
| 119 : | mov edx, ecx | ||
| 120 : | |||
| 121 : | .y_inner_loop: | ||
| 122 : | prefetchnta [esi + 64] ; non temporal prefetch | ||
| 123 : | prefetchnta [esi + 96] | ||
| 124 : | |||
| 125 : | movq mm1, [esi] ; read from src | ||
| 126 : | movq mm2, [esi + 8] | ||
| 127 : | movq mm3, [esi + 16] | ||
| 128 : | movq mm4, [esi + 24] | ||
| 129 : | movq mm5, [esi + 32] | ||
| 130 : | movq mm6, [esi + 40] | ||
| 131 : | movq mm7, [esi + 48] | ||
| 132 : | movq mm0, [esi + 56] | ||
| 133 : | |||
| 134 : | movntq [edi], mm1 ; write to y_out | ||
| 135 : | movntq [edi + 8], mm2 | ||
| 136 : | movntq [edi + 16], mm3 | ||
| 137 : | movntq [edi + 24], mm4 | ||
| 138 : | movntq [edi + 32], mm5 | ||
| 139 : | movntq [edi + 40], mm6 | ||
| 140 : | movntq [edi + 48], mm7 | ||
| 141 : | movntq [edi + 56], mm0 | ||
| 142 : | |||
| 143 : | add esi, 64 | ||
| 144 : | add edi, 64 | ||
| 145 : | dec ecx | ||
| 146 : | jnz .y_inner_loop | ||
| 147 : | |||
| 148 : | dec ebp | ||
| 149 : | jz .y_outer_loop | ||
| 150 : | |||
| 151 : | .y_remainder_loop: | ||
| 152 : | movq mm1, [esi] ; read from src | ||
| 153 : | movq mm2, [esi + 8] | ||
| 154 : | |||
| 155 : | movntq [edi], mm1 ; write to y_out | ||
| 156 : | movntq [edi + 8], mm2 | ||
| 157 : | |||
| 158 : | add esi, 16 | ||
| 159 : | add edi, 16 | ||
| 160 : | dec ebp | ||
| 161 : | jnz .y_remainder_loop | ||
| 162 : | |||
| 163 : | .y_outer_loop: | ||
| 164 : | mov ebp, [remainder] | ||
| 165 : | mov ecx, edx | ||
| 166 : | add edi, ebx | ||
| 167 : | |||
| 168 : | dec eax | ||
| 169 : | jnz near .y_inner_loop | ||
| 170 : | |||
| 171 : | mov eax, [esp + 40 + localsize] ; height -> eax | ||
| 172 : | mov ebx, [esp + 44 + localsize] ; stride -> ebx | ||
| 173 : | mov ecx, [esp + 36 + localsize] ; width -> ecx | ||
| 174 : | mov edi, [esp + 24 + localsize] ; u_out -> edi | ||
| 175 : | |||
| 176 : | shr ecx, 1 ; width / 2 -> ecx | ||
| 177 : | shr ebx, 1 ; stride / 2 -> ebx | ||
| 178 : | shr eax, 1 ; height / 2 -> eax | ||
| 179 : | |||
| 180 : | sub ebx, ecx ; stride / 2 - width / 2 -> ebx | ||
| 181 : | |||
| 182 : | mov edx, ecx | ||
| 183 : | mov ebp, ecx | ||
| 184 : | shr edx, 6 | ||
| 185 : | mov ecx, edx ; 64 bytes copied per iteration | ||
| 186 : | shl edx, 6 | ||
| 187 : | sub ebp, edx ; remainder -> ebp | ||
| 188 : | shr ebp, 3 ; 8 bytes per iteration | ||
| 189 : | add ebp, 1 | ||
| 190 : | mov [remainder], ebp | ||
| 191 : | |||
| 192 : | mov edx, ecx | ||
| 193 : | |||
| 194 : | .u_inner_loop: | ||
| 195 : | prefetchnta [esi + 64] ; non temporal prefetch | ||
| 196 : | prefetchnta [esi + 96] | ||
| 197 : | |||
| 198 : | movq mm1, [esi] ; read from src | ||
| 199 : | movq mm2, [esi + 8] | ||
| 200 : | movq mm3, [esi + 16] | ||
| 201 : | movq mm4, [esi + 24] | ||
| 202 : | movq mm5, [esi + 32] | ||
| 203 : | movq mm6, [esi + 40] | ||
| 204 : | movq mm7, [esi + 48] | ||
| 205 : | movq mm0, [esi + 56] | ||
| 206 : | |||
| 207 : | movntq [edi], mm1 ; write to u_out | ||
| 208 : | movntq [edi + 8], mm2 | ||
| 209 : | movntq [edi + 16], mm3 | ||
| 210 : | movntq [edi + 24], mm4 | ||
| 211 : | movntq [edi + 32], mm5 | ||
| 212 : | movntq [edi + 40], mm6 | ||
| 213 : | movntq [edi + 48], mm7 | ||
| 214 : | movntq [edi + 56], mm0 | ||
| 215 : | |||
| 216 : | |||
| 217 : | add esi, 64 | ||
| 218 : | add edi, 64 | ||
| 219 : | dec ecx | ||
| 220 : | jnz .u_inner_loop | ||
| 221 : | |||
| 222 : | dec ebp | ||
| 223 : | jz .u_outer_loop | ||
| 224 : | |||
| 225 : | .u_remainder_loop: | ||
| 226 : | movq mm1, [esi] ; read from src | ||
| 227 : | movntq [edi], mm1 ; write to y_out | ||
| 228 : | |||
| 229 : | add esi, 8 | ||
| 230 : | add edi, 8 | ||
| 231 : | dec ebp | ||
| 232 : | jnz .u_remainder_loop | ||
| 233 : | |||
| 234 : | .u_outer_loop: | ||
| 235 : | mov ebp, [remainder] | ||
| 236 : | mov ecx, edx | ||
| 237 : | add edi, ebx | ||
| 238 : | |||
| 239 : | dec eax | ||
| 240 : | jnz .u_inner_loop | ||
| 241 : | |||
| 242 : | mov eax, [esp + 40 + localsize] ; height -> eax | ||
| 243 : | mov ecx, [esp + 36 + localsize] ; width -> ecx | ||
| 244 : | mov edi, [esp + 28 + localsize] ; v_out -> edi | ||
| 245 : | |||
| 246 : | shr ecx, 1 ; width / 2 -> ecx | ||
| 247 : | shr eax, 1 ; height / 2 -> eax | ||
| 248 : | |||
| 249 : | mov edx, ecx | ||
| 250 : | mov ebp, ecx | ||
| 251 : | shr edx, 6 | ||
| 252 : | mov ecx, edx ; 64 bytes copied per iteration | ||
| 253 : | shl edx, 6 | ||
| 254 : | sub ebp, edx ; remainder -> ebp | ||
| 255 : | shr ebp, 3 ; 8 bytes per iteration | ||
| 256 : | add ebp, 1 | ||
| 257 : | mov [remainder], ebp | ||
| 258 : | |||
| 259 : | mov edx, ecx | ||
| 260 : | |||
| 261 : | .v_inner_loop: | ||
| 262 : | prefetchnta [esi + 64] ; non temporal prefetch | ||
| 263 : | prefetchnta [esi + 96] | ||
| 264 : | |||
| 265 : | movq mm1, [esi] ; read from src | ||
| 266 : | movq mm2, [esi + 8] | ||
| 267 : | movq mm3, [esi + 16] | ||
| 268 : | movq mm4, [esi + 24] | ||
| 269 : | movq mm5, [esi + 32] | ||
| 270 : | movq mm6, [esi + 40] | ||
| 271 : | movq mm7, [esi + 48] | ||
| 272 : | movq mm0, [esi + 56] | ||
| 273 : | |||
| 274 : | movntq [edi], mm1 ; write to u_out | ||
| 275 : | movntq [edi + 8], mm2 | ||
| 276 : | movntq [edi + 16], mm3 | ||
| 277 : | movntq [edi + 24], mm4 | ||
| 278 : | movntq [edi + 32], mm5 | ||
| 279 : | movntq [edi + 40], mm6 | ||
| 280 : | movntq [edi + 48], mm7 | ||
| 281 : | movntq [edi + 56], mm0 | ||
| 282 : | |||
| 283 : | |||
| 284 : | add esi, 64 | ||
| 285 : | add edi, 64 | ||
| 286 : | dec ecx | ||
| 287 : | jnz .v_inner_loop | ||
| 288 : | |||
| 289 : | dec ebp | ||
| 290 : | jz .v_outer_loop | ||
| 291 : | |||
| 292 : | .v_remainder_loop: | ||
| 293 : | movq mm1, [esi] ; read from src | ||
| 294 : | movntq [edi], mm1 ; write to y_out | ||
| 295 : | |||
| 296 : | add esi, 8 | ||
| 297 : | add edi, 8 | ||
| 298 : | dec ebp | ||
| 299 : | jnz .v_remainder_loop | ||
| 300 : | |||
| 301 : | .v_outer_loop: | ||
| 302 : | mov ebp, [remainder] | ||
| 303 : | mov ecx, edx | ||
| 304 : | add edi, ebx | ||
| 305 : | |||
| 306 : | dec eax | ||
| 307 : | jnz .v_inner_loop | ||
| 308 : | |||
| 309 : | ; local vars deallocation | ||
| 310 : | add esp, localsize | ||
| 311 : | %undef localsize | ||
| 312 : | %undef remainder | ||
| 313 : | |||
| 314 : | pop ebp | ||
| 315 : | pop edi | ||
| 316 : | pop esi | ||
| 317 : | pop ebx | ||
| 318 : | |||
| 319 : | emms | ||
| 320 : | |||
| 321 : | ret | ||
| 322 : | |||
| 323 : | |||
| 324 : | |||
| 325 : | ;------------------------------------------------------------------------------ | ||
| 326 : | ; | ||
| 327 : | ; void yuv_to_yv12_mmx(uint8_t *y_out, | ||
| 328 : | ; uint8_t *u_out, | ||
| 329 : | ; uint8_t *v_out, | ||
| 330 : | ; uint8_t *src, | ||
| 331 : | ; int width, int height, int stride); | ||
| 332 : | ; | ||
| 333 : | ; Attention: This code assumes that width is a multiple of 16 | ||
| 334 : | ; | ||
| 335 : | ;------------------------------------------------------------------------------ | ||
| 336 : | |||
| 337 : | cglobal yuv_to_yv12_mmx | ||
| 338 : | yuv_to_yv12_mmx: | ||
| 339 : | |||
| 340 : | push ebx | ||
| 341 : | push esi | ||
| 342 : | push edi | ||
| 343 : | push ebp | ||
| 344 : | |||
| 345 : | ; local vars allocation | ||
| 346 : | %define localsize 4 | ||
| 347 : | %define remainder esp | ||
| 348 : | sub esp, localsize | ||
| 349 : | |||
| 350 : | |||
| 351 : | ; function code | ||
| 352 : | mov eax, [esp + 40 + localsize] ; height -> eax | ||
| 353 : | mov ebx, [esp + 44 + localsize] ; stride -> ebx | ||
| 354 : | mov esi, [esp + 32 + localsize] ; src -> esi | ||
| 355 : | mov edi, [esp + 20 + localsize] ; y_out -> edi | ||
| 356 : | edgomez | 331 | mov ecx, [esp + 36 + localsize] ; width -> ecx |
| 357 : | edgomez | 328 | |
| 358 : | sub ebx, ecx ; stride - width -> ebx | ||
| 359 : | |||
| 360 : | mov edx, ecx | ||
| 361 : | mov ebp, ecx | ||
| 362 : | shr edx, 6 | ||
| 363 : | mov ecx, edx ; 64 bytes copied per iteration | ||
| 364 : | shl edx, 6 | ||
| 365 : | sub ebp, edx ; mainder -> ebp | ||
| 366 : | shr ebp, 4 ; 16 bytes per iteration | ||
| 367 : | add ebp, 1 | ||
| 368 : | mov [remainder], ebp | ||
| 369 : | |||
| 370 : | mov edx, ecx | ||
| 371 : | |||
| 372 : | .y_inner_loop: | ||
| 373 : | movq mm1, [esi] ; read from src | ||
| 374 : | movq mm2, [esi + 8] | ||
| 375 : | movq mm3, [esi + 16] | ||
| 376 : | movq mm4, [esi + 24] | ||
| 377 : | movq mm5, [esi + 32] | ||
| 378 : | movq mm6, [esi + 40] | ||
| 379 : | movq mm7, [esi + 48] | ||
| 380 : | movq mm0, [esi + 56] | ||
| 381 : | |||
| 382 : | movq [edi], mm1 ; write to y_out | ||
| 383 : | movq [edi + 8], mm2 | ||
| 384 : | movq [edi + 16], mm3 | ||
| 385 : | movq [edi + 24], mm4 | ||
| 386 : | movq [edi + 32], mm5 | ||
| 387 : | movq [edi + 40], mm6 | ||
| 388 : | movq [edi + 48], mm7 | ||
| 389 : | movq [edi + 56], mm0 | ||
| 390 : | |||
| 391 : | add esi, 64 | ||
| 392 : | add edi, 64 | ||
| 393 : | dec ecx | ||
| 394 : | jnz .y_inner_loop | ||
| 395 : | |||
| 396 : | dec ebp | ||
| 397 : | jz .y_outer_loop | ||
| 398 : | |||
| 399 : | .y_remainder_loop: | ||
| 400 : | movq mm1, [esi] ; read from src | ||
| 401 : | movq mm2, [esi + 8] | ||
| 402 : | |||
| 403 : | movq [edi], mm1 ; write to y_out | ||
| 404 : | movq [edi + 8], mm2 | ||
| 405 : | |||
| 406 : | add esi, 16 | ||
| 407 : | add edi, 16 | ||
| 408 : | dec ebp | ||
| 409 : | jnz .y_remainder_loop | ||
| 410 : | |||
| 411 : | .y_outer_loop: | ||
| 412 : | mov ebp, [remainder] | ||
| 413 : | mov ecx, edx | ||
| 414 : | add edi, ebx | ||
| 415 : | |||
| 416 : | dec eax | ||
| 417 : | jnz near .y_inner_loop | ||
| 418 : | |||
| 419 : | mov eax, [esp + 40 + localsize] ; height -> eax | ||
| 420 : | mov ebx, [esp + 44 + localsize] ; stride -> ebx | ||
| 421 : | mov ecx, [esp + 36 + localsize] ; width -> ecx | ||
| 422 : | mov edi, [esp + 24 + localsize] ; u_out -> edi | ||
| 423 : | |||
| 424 : | shr ecx, 1 ; width / 2 -> ecx | ||
| 425 : | shr ebx, 1 ; stride / 2 -> ebx | ||
| 426 : | shr eax, 1 ; height / 2 -> eax | ||
| 427 : | |||
| 428 : | sub ebx, ecx ; stride / 2 - width / 2 -> ebx | ||
| 429 : | |||
| 430 : | mov edx, ecx | ||
| 431 : | mov ebp, ecx | ||
| 432 : | shr edx, 6 | ||
| 433 : | mov ecx, edx ; 64 bytes copied per iteration | ||
| 434 : | shl edx, 6 | ||
| 435 : | sub ebp, edx ; remainder -> ebp | ||
| 436 : | shr ebp, 3 ; 8 bytes per iteration | ||
| 437 : | add ebp, 1 | ||
| 438 : | mov [remainder], ebp | ||
| 439 : | |||
| 440 : | mov edx, ecx | ||
| 441 : | |||
| 442 : | .u_inner_loop: | ||
| 443 : | movq mm1, [esi] ; read from src | ||
| 444 : | movq mm2, [esi + 8] | ||
| 445 : | movq mm3, [esi + 16] | ||
| 446 : | movq mm4, [esi + 24] | ||
| 447 : | movq mm5, [esi + 32] | ||
| 448 : | movq mm6, [esi + 40] | ||
| 449 : | movq mm7, [esi + 48] | ||
| 450 : | movq mm0, [esi + 56] | ||
| 451 : | |||
| 452 : | movq [edi], mm1 ; write to u_out | ||
| 453 : | movq [edi + 8], mm2 | ||
| 454 : | movq [edi + 16], mm3 | ||
| 455 : | movq [edi + 24], mm4 | ||
| 456 : | movq [edi + 32], mm5 | ||
| 457 : | movq [edi + 40], mm6 | ||
| 458 : | movq [edi + 48], mm7 | ||
| 459 : | movq [edi + 56], mm0 | ||
| 460 : | |||
| 461 : | |||
| 462 : | add esi, 64 | ||
| 463 : | add edi, 64 | ||
| 464 : | dec ecx | ||
| 465 : | jnz .u_inner_loop | ||
| 466 : | |||
| 467 : | dec ebp | ||
| 468 : | jz .u_outer_loop | ||
| 469 : | |||
| 470 : | .u_remainder_loop: | ||
| 471 : | movq mm1, [esi] ; read from src | ||
| 472 : | movq [edi], mm1 ; write to y_out | ||
| 473 : | |||
| 474 : | add esi, 8 | ||
| 475 : | add edi, 8 | ||
| 476 : | dec ebp | ||
| 477 : | jnz .u_remainder_loop | ||
| 478 : | |||
| 479 : | .u_outer_loop: | ||
| 480 : | mov ebp, [remainder] | ||
| 481 : | mov ecx, edx | ||
| 482 : | add edi, ebx | ||
| 483 : | |||
| 484 : | dec eax | ||
| 485 : | jnz .u_inner_loop | ||
| 486 : | |||
| 487 : | mov eax, [esp + 40 + localsize] ; height -> eax | ||
| 488 : | mov ecx, [esp + 36 + localsize] ; width -> ecx | ||
| 489 : | mov edi, [esp + 28 + localsize] ; v_out -> edi | ||
| 490 : | |||
| 491 : | shr ecx, 1 ; width / 2 -> ecx | ||
| 492 : | shr eax, 1 ; height / 2 -> eax | ||
| 493 : | |||
| 494 : | mov edx, ecx | ||
| 495 : | mov ebp, ecx | ||
| 496 : | shr edx, 6 | ||
| 497 : | mov ecx, edx ; 64 bytes copied per iteration | ||
| 498 : | shl edx, 6 | ||
| 499 : | sub ebp, edx ; remainder -> ebp | ||
| 500 : | shr ebp, 3 ; 8 bytes per iteration | ||
| 501 : | add ebp, 1 | ||
| 502 : | mov [remainder], ebp | ||
| 503 : | |||
| 504 : | mov edx, ecx | ||
| 505 : | |||
| 506 : | .v_inner_loop: | ||
| 507 : | movq mm1, [esi] ; read from src | ||
| 508 : | movq mm2, [esi + 8] | ||
| 509 : | movq mm3, [esi + 16] | ||
| 510 : | movq mm4, [esi + 24] | ||
| 511 : | movq mm5, [esi + 32] | ||
| 512 : | movq mm6, [esi + 40] | ||
| 513 : | movq mm7, [esi + 48] | ||
| 514 : | movq mm0, [esi + 56] | ||
| 515 : | |||
| 516 : | movq [edi], mm1 ; write to u_out | ||
| 517 : | movq [edi + 8], mm2 | ||
| 518 : | movq [edi + 16], mm3 | ||
| 519 : | movq [edi + 24], mm4 | ||
| 520 : | movq [edi + 32], mm5 | ||
| 521 : | movq [edi + 40], mm6 | ||
| 522 : | movq [edi + 48], mm7 | ||
| 523 : | movq [edi + 56], mm0 | ||
| 524 : | |||
| 525 : | |||
| 526 : | add esi, 64 | ||
| 527 : | add edi, 64 | ||
| 528 : | dec ecx | ||
| 529 : | jnz .v_inner_loop | ||
| 530 : | |||
| 531 : | dec ebp | ||
| 532 : | jz .v_outer_loop | ||
| 533 : | |||
| 534 : | .v_remainder_loop: | ||
| 535 : | movq mm1, [esi] ; read from src | ||
| 536 : | movq [edi], mm1 ; write to y_out | ||
| 537 : | |||
| 538 : | add esi, 8 | ||
| 539 : | add edi, 8 | ||
| 540 : | dec ebp | ||
| 541 : | jnz .v_remainder_loop | ||
| 542 : | |||
| 543 : | .v_outer_loop: | ||
| 544 : | mov ebp, [remainder] | ||
| 545 : | mov ecx, edx | ||
| 546 : | add edi, ebx | ||
| 547 : | |||
| 548 : | dec eax | ||
| 549 : | jnz .v_inner_loop | ||
| 550 : | |||
| 551 : | ; local vars deallocation | ||
| 552 : | add esp, localsize | ||
| 553 : | %undef localsize | ||
| 554 : | %undef remainder | ||
| 555 : | |||
| 556 : | pop ebp | ||
| 557 : | pop edi | ||
| 558 : | pop esi | ||
| 559 : | pop ebx | ||
| 560 : | |||
| 561 : | emms | ||
| 562 : | |||
| 563 : | ret |
| No admin address has been configured | ViewVC Help |
| Powered by ViewVC 1.0.4 |