Parent Directory
| 
Revision Log
Revision 430 - (view) (download)
| 1 : | chl | 430 | ;/***************************************************************************** |
| 2 : | Isibaar | 262 | ; * |
| 3 : | chl | 430 | ; * XVID MPEG-4 VIDEO CODEC |
| 4 : | ; * xmm (extended mmx) sum of absolute difference | ||
| 5 : | Isibaar | 262 | ; * |
| 6 : | chl | 430 | ; * Copyright(C) 2002 Peter Ross <pross@xvid.org> |
| 7 : | ; * Copyright(C) 2002 Michael Militzer <michael@xvid.org> | ||
| 8 : | ; * Copyright(C) 2002 -Skal- | ||
| 9 : | Isibaar | 262 | ; * |
| 10 : | chl | 430 | ; * This program is an implementation of a part of one or more MPEG-4 |
| 11 : | ; * Video tools as specified in ISO/IEC 14496-2 standard. Those intending | ||
| 12 : | ; * to use this software module in hardware or software products are | ||
| 13 : | ; * advised that its use may infringe existing patents or copyrights, and | ||
| 14 : | ; * any such use would be at such party's own risk. The original | ||
| 15 : | ; * developer of this software module and his/her company, and subsequent | ||
| 16 : | ; * editors and their companies, will have no liability for use of this | ||
| 17 : | ; * software or modifications or derivatives thereof. | ||
| 18 : | Isibaar | 262 | ; * |
| 19 : | chl | 430 | ; * This program is free software; you can redistribute it and/or modify |
| 20 : | ; * it under the terms of the GNU General Public License as published by | ||
| 21 : | ; * the Free Software Foundation; either version 2 of the License, or | ||
| 22 : | ; * (at your option) any later version. | ||
| 23 : | Isibaar | 262 | ; * |
| 24 : | chl | 430 | ; * This program is distributed in the hope that it will be useful, |
| 25 : | ; * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
| 26 : | ; * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
| 27 : | ; * GNU General Public License for more details. | ||
| 28 : | Isibaar | 262 | ; * |
| 29 : | chl | 430 | ; * You should have received a copy of the GNU General Public License |
| 30 : | ; * along with this program; if not, write to the Free Software | ||
| 31 : | ; * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | ||
| 32 : | Isibaar | 262 | ; * |
| 33 : | ; *************************************************************************/ | ||
| 34 : | |||
| 35 : | bits 32 | ||
| 36 : | suxen_drol | 329 | |
| 37 : | Isibaar | 262 | %macro cglobal 1 |
| 38 : | %ifdef PREFIX | ||
| 39 : | global _%1 | ||
| 40 : | %define %1 _%1 | ||
| 41 : | %else | ||
| 42 : | global %1 | ||
| 43 : | %endif | ||
| 44 : | %endmacro | ||
| 45 : | |||
| 46 : | section .data | ||
| 47 : | |||
| 48 : | align 16 | ||
| 49 : | mmx_one times 4 dw 1 | ||
| 50 : | |||
| 51 : | section .text | ||
| 52 : | |||
| 53 : | cglobal sad16_xmm | ||
| 54 : | cglobal sad8_xmm | ||
| 55 : | cglobal sad16bi_xmm | ||
| 56 : | cglobal sad8bi_xmm | ||
| 57 : | cglobal dev16_xmm | ||
| 58 : | |||
| 59 : | ;=========================================================================== | ||
| 60 : | ; | ||
| 61 : | ; uint32_t sad16_xmm(const uint8_t * const cur, | ||
| 62 : | ; const uint8_t * const ref, | ||
| 63 : | suxen_drol | 329 | ; const uint32_t stride, |
| 64 : | Isibaar | 262 | ; const uint32_t best_sad); |
| 65 : | suxen_drol | 329 | ; |
| 66 : | Isibaar | 262 | ;=========================================================================== |
| 67 : | |||
| 68 : | %macro SAD_16x16_SSE 0 | ||
| 69 : | movq mm0, [eax] | ||
| 70 : | psadbw mm0, [edx] | ||
| 71 : | movq mm1, [eax+8] | ||
| 72 : | add eax, ecx | ||
| 73 : | psadbw mm1, [edx+8] | ||
| 74 : | paddusw mm5,mm0 | ||
| 75 : | add edx, ecx | ||
| 76 : | paddusw mm6,mm1 | ||
| 77 : | %endmacro | ||
| 78 : | |||
| 79 : | align 16 | ||
| 80 : | sad16_xmm: | ||
| 81 : | |||
| 82 : | mov eax, [esp+ 4] ; Src1 | ||
| 83 : | mov edx, [esp+ 8] ; Src2 | ||
| 84 : | mov ecx, [esp+12] ; Stride | ||
| 85 : | |||
| 86 : | pxor mm5, mm5 ; accum1 | ||
| 87 : | pxor mm6, mm6 ; accum2 | ||
| 88 : | |||
| 89 : | SAD_16x16_SSE | ||
| 90 : | SAD_16x16_SSE | ||
| 91 : | SAD_16x16_SSE | ||
| 92 : | SAD_16x16_SSE | ||
| 93 : | SAD_16x16_SSE | ||
| 94 : | SAD_16x16_SSE | ||
| 95 : | SAD_16x16_SSE | ||
| 96 : | SAD_16x16_SSE | ||
| 97 : | |||
| 98 : | SAD_16x16_SSE | ||
| 99 : | SAD_16x16_SSE | ||
| 100 : | SAD_16x16_SSE | ||
| 101 : | SAD_16x16_SSE | ||
| 102 : | SAD_16x16_SSE | ||
| 103 : | SAD_16x16_SSE | ||
| 104 : | SAD_16x16_SSE | ||
| 105 : | SAD_16x16_SSE | ||
| 106 : | |||
| 107 : | paddusw mm6,mm5 | ||
| 108 : | movd eax, mm6 | ||
| 109 : | ret | ||
| 110 : | |||
| 111 : | |||
| 112 : | ;=========================================================================== | ||
| 113 : | ; | ||
| 114 : | ; uint32_t sad8_xmm(const uint8_t * const cur, | ||
| 115 : | ; const uint8_t * const ref, | ||
| 116 : | ; const uint32_t stride); | ||
| 117 : | ; | ||
| 118 : | ;=========================================================================== | ||
| 119 : | |||
| 120 : | suxen_drol | 329 | %macro SAD_8x8_SSE 0 |
| 121 : | movq mm0, [eax] | ||
| 122 : | movq mm1, [eax+ecx] | ||
| 123 : | |||
| 124 : | psadbw mm0, [edx] | ||
| 125 : | psadbw mm1, [edx+ecx] | ||
| 126 : | add eax, ebx | ||
| 127 : | add edx, ebx | ||
| 128 : | |||
| 129 : | paddusw mm5,mm0 | ||
| 130 : | paddusw mm6,mm1 | ||
| 131 : | %endmacro | ||
| 132 : | |||
| 133 : | align 16 | ||
| 134 : | sad8_xmm: | ||
| 135 : | |||
| 136 : | mov eax, [esp+ 4] ; Src1 | ||
| 137 : | mov edx, [esp+ 8] ; Src2 | ||
| 138 : | mov ecx, [esp+12] ; Stride | ||
| 139 : | push ebx | ||
| 140 : | lea ebx, [ecx+ecx] | ||
| 141 : | |||
| 142 : | pxor mm5, mm5 ; accum1 | ||
| 143 : | pxor mm6, mm6 ; accum2 | ||
| 144 : | |||
| 145 : | SAD_8x8_SSE | ||
| 146 : | SAD_8x8_SSE | ||
| 147 : | SAD_8x8_SSE | ||
| 148 : | |||
| 149 : | movq mm0, [eax] | ||
| 150 : | movq mm1, [eax+ecx] | ||
| 151 : | psadbw mm0, [edx] | ||
| 152 : | psadbw mm1, [edx+ecx] | ||
| 153 : | |||
| 154 : | pop ebx | ||
| 155 : | |||
| 156 : | paddusw mm5,mm0 | ||
| 157 : | paddusw mm6,mm1 | ||
| 158 : | |||
| 159 : | paddusw mm6,mm5 | ||
| 160 : | movd eax, mm6 | ||
| 161 : | |||
| 162 : | ret | ||
| 163 : | |||
| 164 : | |||
| 165 : | ;=========================================================================== | ||
| 166 : | ; | ||
| 167 : | ; uint32_t sad16bi_xmm(const uint8_t * const cur, | ||
| 168 : | ; const uint8_t * const ref1, | ||
| 169 : | ; const uint8_t * const ref2, | ||
| 170 : | ; const uint32_t stride); | ||
| 171 : | ; | ||
| 172 : | ;=========================================================================== | ||
| 173 : | |||
| 174 : | Isibaar | 262 | %macro SADBI_16x16_SSE 0 |
| 175 : | movq mm0, [eax] | ||
| 176 : | movq mm1, [eax+8] | ||
| 177 : | |||
| 178 : | movq mm2, [edx] | ||
| 179 : | movq mm3, [edx+8] | ||
| 180 : | |||
| 181 : | pavgb mm2, [ebx] | ||
| 182 : | add edx, ecx | ||
| 183 : | |||
| 184 : | pavgb mm3, [ebx+8] | ||
| 185 : | add ebx, ecx | ||
| 186 : | |||
| 187 : | psadbw mm0, mm2 | ||
| 188 : | add eax, ecx | ||
| 189 : | |||
| 190 : | psadbw mm1, mm3 | ||
| 191 : | paddusw mm5,mm0 | ||
| 192 : | |||
| 193 : | paddusw mm6,mm1 | ||
| 194 : | %endmacro | ||
| 195 : | |||
| 196 : | align 16 | ||
| 197 : | sad16bi_xmm: | ||
| 198 : | push ebx | ||
| 199 : | mov eax, [esp+4+ 4] ; Src | ||
| 200 : | mov edx, [esp+4+ 8] ; Ref1 | ||
| 201 : | mov ebx, [esp+4+12] ; Ref2 | ||
| 202 : | mov ecx, [esp+4+16] ; Stride | ||
| 203 : | |||
| 204 : | pxor mm5, mm5 ; accum1 | ||
| 205 : | pxor mm6, mm6 ; accum2 | ||
| 206 : | |||
| 207 : | SADBI_16x16_SSE | ||
| 208 : | SADBI_16x16_SSE | ||
| 209 : | SADBI_16x16_SSE | ||
| 210 : | SADBI_16x16_SSE | ||
| 211 : | SADBI_16x16_SSE | ||
| 212 : | SADBI_16x16_SSE | ||
| 213 : | SADBI_16x16_SSE | ||
| 214 : | SADBI_16x16_SSE | ||
| 215 : | |||
| 216 : | SADBI_16x16_SSE | ||
| 217 : | SADBI_16x16_SSE | ||
| 218 : | SADBI_16x16_SSE | ||
| 219 : | SADBI_16x16_SSE | ||
| 220 : | SADBI_16x16_SSE | ||
| 221 : | SADBI_16x16_SSE | ||
| 222 : | SADBI_16x16_SSE | ||
| 223 : | SADBI_16x16_SSE | ||
| 224 : | |||
| 225 : | paddusw mm6,mm5 | ||
| 226 : | movd eax, mm6 | ||
| 227 : | pop ebx | ||
| 228 : | ret | ||
| 229 : | |||
| 230 : | ;=========================================================================== | ||
| 231 : | ; | ||
| 232 : | ; uint32_t sad8bi_xmm(const uint8_t * const cur, | ||
| 233 : | ; const uint8_t * const ref1, | ||
| 234 : | ; const uint8_t * const ref2, | ||
| 235 : | ; const uint32_t stride); | ||
| 236 : | ; | ||
| 237 : | ;=========================================================================== | ||
| 238 : | |||
| 239 : | suxen_drol | 329 | %macro SADBI_8x8_XMM 0 |
| 240 : | movq mm0, [eax] | ||
| 241 : | movq mm1, [eax+ecx] | ||
| 242 : | Isibaar | 262 | |
| 243 : | suxen_drol | 329 | movq mm2, [edx] |
| 244 : | movq mm3, [edx+ecx] | ||
| 245 : | Isibaar | 262 | |
| 246 : | suxen_drol | 329 | pavgb mm2, [ebx] |
| 247 : | lea edx, [edx+2*ecx] | ||
| 248 : | Isibaar | 262 | |
| 249 : | suxen_drol | 329 | pavgb mm3, [ebx+ecx] |
| 250 : | lea ebx, [ebx+2*ecx] | ||
| 251 : | Isibaar | 262 | |
| 252 : | suxen_drol | 329 | psadbw mm0, mm2 |
| 253 : | lea eax, [eax+2*ecx] | ||
| 254 : | Isibaar | 262 | |
| 255 : | suxen_drol | 329 | psadbw mm1, mm3 |
| 256 : | paddusw mm5,mm0 | ||
| 257 : | Isibaar | 262 | |
| 258 : | suxen_drol | 329 | paddusw mm6,mm1 |
| 259 : | %endmacro | ||
| 260 : | Isibaar | 262 | |
| 261 : | suxen_drol | 329 | align 16 |
| 262 : | sad8bi_xmm: | ||
| 263 : | push ebx | ||
| 264 : | mov eax, [esp+4+ 4] ; Src | ||
| 265 : | mov edx, [esp+4+ 8] ; Ref1 | ||
| 266 : | mov ebx, [esp+4+12] ; Ref2 | ||
| 267 : | mov ecx, [esp+4+16] ; Stride | ||
| 268 : | Isibaar | 262 | |
| 269 : | suxen_drol | 329 | pxor mm5, mm5 ; accum1 |
| 270 : | pxor mm6, mm6 ; accum2 | ||
| 271 : | .Loop | ||
| 272 : | SADBI_8x8_XMM | ||
| 273 : | SADBI_8x8_XMM | ||
| 274 : | SADBI_8x8_XMM | ||
| 275 : | SADBI_8x8_XMM | ||
| 276 : | Isibaar | 262 | |
| 277 : | suxen_drol | 329 | paddusw mm6,mm5 |
| 278 : | movd eax, mm6 | ||
| 279 : | pop ebx | ||
| 280 : | ret | ||
| 281 : | Isibaar | 262 | |
| 282 : | suxen_drol | 329 | |
| 283 : | ;=========================================================================== | ||
| 284 : | ; | ||
| 285 : | ; uint32_t dev16_xmm(const uint8_t * const cur, | ||
| 286 : | ; const uint32_t stride); | ||
| 287 : | ; | ||
| 288 : | ;=========================================================================== | ||
| 289 : | |||
| 290 : | %macro MEAN_16x16_SSE 0 | ||
| 291 : | Isibaar | 262 | movq mm0, [eax] |
| 292 : | movq mm1, [eax+8] | ||
| 293 : | psadbw mm0, mm7 | ||
| 294 : | psadbw mm1, mm7 | ||
| 295 : | add eax, ecx | ||
| 296 : | paddw mm5, mm0 | ||
| 297 : | paddw mm6, mm1 | ||
| 298 : | %endmacro | ||
| 299 : | |||
| 300 : | %macro ABS_16x16_SSE 0 | ||
| 301 : | movq mm0, [eax] | ||
| 302 : | movq mm1, [eax+8] | ||
| 303 : | psadbw mm0, mm4 | ||
| 304 : | psadbw mm1, mm4 | ||
| 305 : | lea eax,[eax+ecx] | ||
| 306 : | paddw mm5, mm0 | ||
| 307 : | paddw mm6, mm1 | ||
| 308 : | %endmacro | ||
| 309 : | |||
| 310 : | align 16 | ||
| 311 : | dev16_xmm: | ||
| 312 : | |||
| 313 : | mov eax, [esp+ 4] ; Src | ||
| 314 : | mov ecx, [esp+ 8] ; Stride | ||
| 315 : | |||
| 316 : | pxor mm7, mm7 ; zero | ||
| 317 : | pxor mm5, mm5 ; mean accums | ||
| 318 : | pxor mm6, mm6 | ||
| 319 : | |||
| 320 : | MEAN_16x16_SSE | ||
| 321 : | MEAN_16x16_SSE | ||
| 322 : | MEAN_16x16_SSE | ||
| 323 : | MEAN_16x16_SSE | ||
| 324 : | MEAN_16x16_SSE | ||
| 325 : | MEAN_16x16_SSE | ||
| 326 : | MEAN_16x16_SSE | ||
| 327 : | MEAN_16x16_SSE | ||
| 328 : | |||
| 329 : | MEAN_16x16_SSE | ||
| 330 : | MEAN_16x16_SSE | ||
| 331 : | MEAN_16x16_SSE | ||
| 332 : | MEAN_16x16_SSE | ||
| 333 : | MEAN_16x16_SSE | ||
| 334 : | MEAN_16x16_SSE | ||
| 335 : | MEAN_16x16_SSE | ||
| 336 : | MEAN_16x16_SSE | ||
| 337 : | |||
| 338 : | paddusw mm6, mm5 | ||
| 339 : | |||
| 340 : | movq mm4, mm6 | ||
| 341 : | psllq mm4, 32 | ||
| 342 : | paddd mm4, mm6 | ||
| 343 : | psrld mm4, 8 ; /= (16*16) | ||
| 344 : | |||
| 345 : | packssdw mm4, mm4 | ||
| 346 : | packuswb mm4, mm4 | ||
| 347 : | |||
| 348 : | ; mm4 contains the mean | ||
| 349 : | |||
| 350 : | mov eax, [esp+ 4] ; Src | ||
| 351 : | |||
| 352 : | pxor mm5, mm5 ; sums | ||
| 353 : | pxor mm6, mm6 | ||
| 354 : | |||
| 355 : | ABS_16x16_SSE | ||
| 356 : | ABS_16x16_SSE | ||
| 357 : | ABS_16x16_SSE | ||
| 358 : | ABS_16x16_SSE | ||
| 359 : | ABS_16x16_SSE | ||
| 360 : | ABS_16x16_SSE | ||
| 361 : | ABS_16x16_SSE | ||
| 362 : | ABS_16x16_SSE | ||
| 363 : | |||
| 364 : | ABS_16x16_SSE | ||
| 365 : | ABS_16x16_SSE | ||
| 366 : | ABS_16x16_SSE | ||
| 367 : | ABS_16x16_SSE | ||
| 368 : | ABS_16x16_SSE | ||
| 369 : | ABS_16x16_SSE | ||
| 370 : | ABS_16x16_SSE | ||
| 371 : | ABS_16x16_SSE | ||
| 372 : | |||
| 373 : | paddusw mm6, mm5 | ||
| 374 : | movq mm7, mm6 | ||
| 375 : | psllq mm7, 32 | ||
| 376 : | paddd mm6, mm7 | ||
| 377 : | |||
| 378 : | movd eax, mm6 | ||
| 379 : | chl | 430 | ret |
| No admin address has been configured | ViewVC Help |
| Powered by ViewVC 1.0.4 |