The Tor Browser / annotate

media/libvpx/vp8/common/x86/variance_sse2.c@ac0c01689b40 (annotated)

media/libvpx/vp8/common/x86/variance_sse2.c

Thu, 15 Jan 2015 15:59:08 +0100

author
Michael Schloh von Bennewitz <michael@schloh.com>
date
Thu, 15 Jan 2015 15:59:08 +0100
branch
TOR_BUG_9701
changeset 10
ac0c01689b40
permissions
-rw-r--r--

Implement a real Private Browsing Mode condition by changing the API/ABI;
This solves Tor bug #9701, complying with disk avoidance documented in
https://www.torproject.org/projects/torbrowser/design/#disk-avoidance.

michael@0 1 /*
michael@0 2 * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
michael@0 3 *
michael@0 4 * Use of this source code is governed by a BSD-style license
michael@0 5 * that can be found in the LICENSE file in the root of the source
michael@0 6 * tree. An additional intellectual property rights grant can be found
michael@0 7 * in the file PATENTS. All contributing project authors may
michael@0 8 * be found in the AUTHORS file in the root of the source tree.
michael@0 9 */
michael@0 10
michael@0 11 #include "vpx_config.h"
michael@0 12 #include "vp8/common/variance.h"
michael@0 13 #include "vp8/common/pragmas.h"
michael@0 14 #include "vpx_ports/mem.h"
michael@0 15 #include "vp8/common/x86/filter_x86.h"
michael@0 16
michael@0 17 extern void filter_block1d_h6_mmx(const unsigned char *src_ptr, unsigned short *output_ptr, unsigned int src_pixels_per_line, unsigned int pixel_step, unsigned int output_height, unsigned int output_width, short *filter);
michael@0 18 extern void filter_block1d_v6_mmx(const short *src_ptr, unsigned char *output_ptr, unsigned int pixels_per_line, unsigned int pixel_step, unsigned int output_height, unsigned int output_width, short *filter);
michael@0 19 extern void filter_block1d8_h6_sse2(const unsigned char *src_ptr, unsigned short *output_ptr, unsigned int src_pixels_per_line, unsigned int pixel_step, unsigned int output_height, unsigned int output_width, short *filter);
michael@0 20 extern void filter_block1d8_v6_sse2(const short *src_ptr, unsigned char *output_ptr, unsigned int pixels_per_line, unsigned int pixel_step, unsigned int output_height, unsigned int output_width, short *filter);
michael@0 21
michael@0 22 extern void vp8_filter_block2d_bil4x4_var_mmx
michael@0 23 (
michael@0 24 const unsigned char *ref_ptr,
michael@0 25 int ref_pixels_per_line,
michael@0 26 const unsigned char *src_ptr,
michael@0 27 int src_pixels_per_line,
michael@0 28 const short *HFilter,
michael@0 29 const short *VFilter,
michael@0 30 int *sum,
michael@0 31 unsigned int *sumsquared
michael@0 32 );
michael@0 33
michael@0 34 extern unsigned int vp8_get4x4var_mmx
michael@0 35 (
michael@0 36 const unsigned char *src_ptr,
michael@0 37 int source_stride,
michael@0 38 const unsigned char *ref_ptr,
michael@0 39 int recon_stride,
michael@0 40 unsigned int *SSE,
michael@0 41 int *Sum
michael@0 42 );
michael@0 43
michael@0 44 unsigned int vp8_get_mb_ss_sse2
michael@0 45 (
michael@0 46 const short *src_ptr
michael@0 47 );
michael@0 48 unsigned int vp8_get16x16var_sse2
michael@0 49 (
michael@0 50 const unsigned char *src_ptr,
michael@0 51 int source_stride,
michael@0 52 const unsigned char *ref_ptr,
michael@0 53 int recon_stride,
michael@0 54 unsigned int *SSE,
michael@0 55 int *Sum
michael@0 56 );
michael@0 57 unsigned int vp8_get8x8var_sse2
michael@0 58 (
michael@0 59 const unsigned char *src_ptr,
michael@0 60 int source_stride,
michael@0 61 const unsigned char *ref_ptr,
michael@0 62 int recon_stride,
michael@0 63 unsigned int *SSE,
michael@0 64 int *Sum
michael@0 65 );
michael@0 66 void vp8_filter_block2d_bil_var_sse2
michael@0 67 (
michael@0 68 const unsigned char *ref_ptr,
michael@0 69 int ref_pixels_per_line,
michael@0 70 const unsigned char *src_ptr,
michael@0 71 int src_pixels_per_line,
michael@0 72 unsigned int Height,
michael@0 73 int xoffset,
michael@0 74 int yoffset,
michael@0 75 int *sum,
michael@0 76 unsigned int *sumsquared
michael@0 77 );
michael@0 78 void vp8_half_horiz_vert_variance8x_h_sse2
michael@0 79 (
michael@0 80 const unsigned char *ref_ptr,
michael@0 81 int ref_pixels_per_line,
michael@0 82 const unsigned char *src_ptr,
michael@0 83 int src_pixels_per_line,
michael@0 84 unsigned int Height,
michael@0 85 int *sum,
michael@0 86 unsigned int *sumsquared
michael@0 87 );
michael@0 88 void vp8_half_horiz_vert_variance16x_h_sse2
michael@0 89 (
michael@0 90 const unsigned char *ref_ptr,
michael@0 91 int ref_pixels_per_line,
michael@0 92 const unsigned char *src_ptr,
michael@0 93 int src_pixels_per_line,
michael@0 94 unsigned int Height,
michael@0 95 int *sum,
michael@0 96 unsigned int *sumsquared
michael@0 97 );
michael@0 98 void vp8_half_horiz_variance8x_h_sse2
michael@0 99 (
michael@0 100 const unsigned char *ref_ptr,
michael@0 101 int ref_pixels_per_line,
michael@0 102 const unsigned char *src_ptr,
michael@0 103 int src_pixels_per_line,
michael@0 104 unsigned int Height,
michael@0 105 int *sum,
michael@0 106 unsigned int *sumsquared
michael@0 107 );
michael@0 108 void vp8_half_horiz_variance16x_h_sse2
michael@0 109 (
michael@0 110 const unsigned char *ref_ptr,
michael@0 111 int ref_pixels_per_line,
michael@0 112 const unsigned char *src_ptr,
michael@0 113 int src_pixels_per_line,
michael@0 114 unsigned int Height,
michael@0 115 int *sum,
michael@0 116 unsigned int *sumsquared
michael@0 117 );
michael@0 118 void vp8_half_vert_variance8x_h_sse2
michael@0 119 (
michael@0 120 const unsigned char *ref_ptr,
michael@0 121 int ref_pixels_per_line,
michael@0 122 const unsigned char *src_ptr,
michael@0 123 int src_pixels_per_line,
michael@0 124 unsigned int Height,
michael@0 125 int *sum,
michael@0 126 unsigned int *sumsquared
michael@0 127 );
michael@0 128 void vp8_half_vert_variance16x_h_sse2
michael@0 129 (
michael@0 130 const unsigned char *ref_ptr,
michael@0 131 int ref_pixels_per_line,
michael@0 132 const unsigned char *src_ptr,
michael@0 133 int src_pixels_per_line,
michael@0 134 unsigned int Height,
michael@0 135 int *sum,
michael@0 136 unsigned int *sumsquared
michael@0 137 );
michael@0 138
michael@0 139 unsigned int vp8_variance4x4_wmt(
michael@0 140 const unsigned char *src_ptr,
michael@0 141 int source_stride,
michael@0 142 const unsigned char *ref_ptr,
michael@0 143 int recon_stride,
michael@0 144 unsigned int *sse)
michael@0 145 {
michael@0 146 unsigned int var;
michael@0 147 int avg;
michael@0 148
michael@0 149 vp8_get4x4var_mmx(src_ptr, source_stride, ref_ptr, recon_stride, &var, &avg) ;
michael@0 150 *sse = var;
michael@0 151 return (var - (((unsigned int)avg * avg) >> 4));
michael@0 152
michael@0 153 }
michael@0 154
michael@0 155 unsigned int vp8_variance8x8_wmt
michael@0 156 (
michael@0 157 const unsigned char *src_ptr,
michael@0 158 int source_stride,
michael@0 159 const unsigned char *ref_ptr,
michael@0 160 int recon_stride,
michael@0 161 unsigned int *sse)
michael@0 162 {
michael@0 163 unsigned int var;
michael@0 164 int avg;
michael@0 165
michael@0 166 vp8_get8x8var_sse2(src_ptr, source_stride, ref_ptr, recon_stride, &var, &avg) ;
michael@0 167 *sse = var;
michael@0 168 return (var - (((unsigned int)avg * avg) >> 6));
michael@0 169
michael@0 170 }
michael@0 171
michael@0 172
michael@0 173 unsigned int vp8_variance16x16_wmt
michael@0 174 (
michael@0 175 const unsigned char *src_ptr,
michael@0 176 int source_stride,
michael@0 177 const unsigned char *ref_ptr,
michael@0 178 int recon_stride,
michael@0 179 unsigned int *sse)
michael@0 180 {
michael@0 181 unsigned int sse0;
michael@0 182 int sum0;
michael@0 183
michael@0 184
michael@0 185 vp8_get16x16var_sse2(src_ptr, source_stride, ref_ptr, recon_stride, &sse0, &sum0) ;
michael@0 186 *sse = sse0;
michael@0 187 return (sse0 - (((unsigned int)sum0 * sum0) >> 8));
michael@0 188 }
michael@0 189 unsigned int vp8_mse16x16_wmt(
michael@0 190 const unsigned char *src_ptr,
michael@0 191 int source_stride,
michael@0 192 const unsigned char *ref_ptr,
michael@0 193 int recon_stride,
michael@0 194 unsigned int *sse)
michael@0 195 {
michael@0 196
michael@0 197 unsigned int sse0;
michael@0 198 int sum0;
michael@0 199 vp8_get16x16var_sse2(src_ptr, source_stride, ref_ptr, recon_stride, &sse0, &sum0) ;
michael@0 200 *sse = sse0;
michael@0 201 return sse0;
michael@0 202
michael@0 203 }
michael@0 204
michael@0 205
michael@0 206 unsigned int vp8_variance16x8_wmt
michael@0 207 (
michael@0 208 const unsigned char *src_ptr,
michael@0 209 int source_stride,
michael@0 210 const unsigned char *ref_ptr,
michael@0 211 int recon_stride,
michael@0 212 unsigned int *sse)
michael@0 213 {
michael@0 214 unsigned int sse0, sse1, var;
michael@0 215 int sum0, sum1, avg;
michael@0 216
michael@0 217 vp8_get8x8var_sse2(src_ptr, source_stride, ref_ptr, recon_stride, &sse0, &sum0) ;
michael@0 218 vp8_get8x8var_sse2(src_ptr + 8, source_stride, ref_ptr + 8, recon_stride, &sse1, &sum1);
michael@0 219
michael@0 220 var = sse0 + sse1;
michael@0 221 avg = sum0 + sum1;
michael@0 222 *sse = var;
michael@0 223 return (var - (((unsigned int)avg * avg) >> 7));
michael@0 224
michael@0 225 }
michael@0 226
michael@0 227 unsigned int vp8_variance8x16_wmt
michael@0 228 (
michael@0 229 const unsigned char *src_ptr,
michael@0 230 int source_stride,
michael@0 231 const unsigned char *ref_ptr,
michael@0 232 int recon_stride,
michael@0 233 unsigned int *sse)
michael@0 234 {
michael@0 235 unsigned int sse0, sse1, var;
michael@0 236 int sum0, sum1, avg;
michael@0 237
michael@0 238 vp8_get8x8var_sse2(src_ptr, source_stride, ref_ptr, recon_stride, &sse0, &sum0) ;
michael@0 239 vp8_get8x8var_sse2(src_ptr + 8 * source_stride, source_stride, ref_ptr + 8 * recon_stride, recon_stride, &sse1, &sum1) ;
michael@0 240
michael@0 241 var = sse0 + sse1;
michael@0 242 avg = sum0 + sum1;
michael@0 243 *sse = var;
michael@0 244 return (var - (((unsigned int)avg * avg) >> 7));
michael@0 245
michael@0 246 }
michael@0 247
michael@0 248 unsigned int vp8_sub_pixel_variance4x4_wmt
michael@0 249 (
michael@0 250 const unsigned char *src_ptr,
michael@0 251 int src_pixels_per_line,
michael@0 252 int xoffset,
michael@0 253 int yoffset,
michael@0 254 const unsigned char *dst_ptr,
michael@0 255 int dst_pixels_per_line,
michael@0 256 unsigned int *sse
michael@0 257 )
michael@0 258 {
michael@0 259 int xsum;
michael@0 260 unsigned int xxsum;
michael@0 261 vp8_filter_block2d_bil4x4_var_mmx(
michael@0 262 src_ptr, src_pixels_per_line,
michael@0 263 dst_ptr, dst_pixels_per_line,
michael@0 264 vp8_bilinear_filters_x86_4[xoffset], vp8_bilinear_filters_x86_4[yoffset],
michael@0 265 &xsum, &xxsum
michael@0 266 );
michael@0 267 *sse = xxsum;
michael@0 268 return (xxsum - (((unsigned int)xsum * xsum) >> 4));
michael@0 269 }
michael@0 270
michael@0 271
michael@0 272 unsigned int vp8_sub_pixel_variance8x8_wmt
michael@0 273 (
michael@0 274 const unsigned char *src_ptr,
michael@0 275 int src_pixels_per_line,
michael@0 276 int xoffset,
michael@0 277 int yoffset,
michael@0 278 const unsigned char *dst_ptr,
michael@0 279 int dst_pixels_per_line,
michael@0 280 unsigned int *sse
michael@0 281 )
michael@0 282 {
michael@0 283 int xsum;
michael@0 284 unsigned int xxsum;
michael@0 285
michael@0 286 if (xoffset == 4 && yoffset == 0)
michael@0 287 {
michael@0 288 vp8_half_horiz_variance8x_h_sse2(
michael@0 289 src_ptr, src_pixels_per_line,
michael@0 290 dst_ptr, dst_pixels_per_line, 8,
michael@0 291 &xsum, &xxsum);
michael@0 292 }
michael@0 293 else if (xoffset == 0 && yoffset == 4)
michael@0 294 {
michael@0 295 vp8_half_vert_variance8x_h_sse2(
michael@0 296 src_ptr, src_pixels_per_line,
michael@0 297 dst_ptr, dst_pixels_per_line, 8,
michael@0 298 &xsum, &xxsum);
michael@0 299 }
michael@0 300 else if (xoffset == 4 && yoffset == 4)
michael@0 301 {
michael@0 302 vp8_half_horiz_vert_variance8x_h_sse2(
michael@0 303 src_ptr, src_pixels_per_line,
michael@0 304 dst_ptr, dst_pixels_per_line, 8,
michael@0 305 &xsum, &xxsum);
michael@0 306 }
michael@0 307 else
michael@0 308 {
michael@0 309 vp8_filter_block2d_bil_var_sse2(
michael@0 310 src_ptr, src_pixels_per_line,
michael@0 311 dst_ptr, dst_pixels_per_line, 8,
michael@0 312 xoffset, yoffset,
michael@0 313 &xsum, &xxsum);
michael@0 314 }
michael@0 315
michael@0 316 *sse = xxsum;
michael@0 317 return (xxsum - (((unsigned int)xsum * xsum) >> 6));
michael@0 318 }
michael@0 319
michael@0 320 unsigned int vp8_sub_pixel_variance16x16_wmt
michael@0 321 (
michael@0 322 const unsigned char *src_ptr,
michael@0 323 int src_pixels_per_line,
michael@0 324 int xoffset,
michael@0 325 int yoffset,
michael@0 326 const unsigned char *dst_ptr,
michael@0 327 int dst_pixels_per_line,
michael@0 328 unsigned int *sse
michael@0 329 )
michael@0 330 {
michael@0 331 int xsum0, xsum1;
michael@0 332 unsigned int xxsum0, xxsum1;
michael@0 333
michael@0 334
michael@0 335 /* note we could avoid these if statements if the calling function
michael@0 336 * just called the appropriate functions inside.
michael@0 337 */
michael@0 338 if (xoffset == 4 && yoffset == 0)
michael@0 339 {
michael@0 340 vp8_half_horiz_variance16x_h_sse2(
michael@0 341 src_ptr, src_pixels_per_line,
michael@0 342 dst_ptr, dst_pixels_per_line, 16,
michael@0 343 &xsum0, &xxsum0);
michael@0 344 }
michael@0 345 else if (xoffset == 0 && yoffset == 4)
michael@0 346 {
michael@0 347 vp8_half_vert_variance16x_h_sse2(
michael@0 348 src_ptr, src_pixels_per_line,
michael@0 349 dst_ptr, dst_pixels_per_line, 16,
michael@0 350 &xsum0, &xxsum0);
michael@0 351 }
michael@0 352 else if (xoffset == 4 && yoffset == 4)
michael@0 353 {
michael@0 354 vp8_half_horiz_vert_variance16x_h_sse2(
michael@0 355 src_ptr, src_pixels_per_line,
michael@0 356 dst_ptr, dst_pixels_per_line, 16,
michael@0 357 &xsum0, &xxsum0);
michael@0 358 }
michael@0 359 else
michael@0 360 {
michael@0 361 vp8_filter_block2d_bil_var_sse2(
michael@0 362 src_ptr, src_pixels_per_line,
michael@0 363 dst_ptr, dst_pixels_per_line, 16,
michael@0 364 xoffset, yoffset,
michael@0 365 &xsum0, &xxsum0
michael@0 366 );
michael@0 367
michael@0 368 vp8_filter_block2d_bil_var_sse2(
michael@0 369 src_ptr + 8, src_pixels_per_line,
michael@0 370 dst_ptr + 8, dst_pixels_per_line, 16,
michael@0 371 xoffset, yoffset,
michael@0 372 &xsum1, &xxsum1
michael@0 373 );
michael@0 374 xsum0 += xsum1;
michael@0 375 xxsum0 += xxsum1;
michael@0 376 }
michael@0 377
michael@0 378 *sse = xxsum0;
michael@0 379 return (xxsum0 - (((unsigned int)xsum0 * xsum0) >> 8));
michael@0 380 }
michael@0 381
michael@0 382 unsigned int vp8_sub_pixel_mse16x16_wmt(
michael@0 383 const unsigned char *src_ptr,
michael@0 384 int src_pixels_per_line,
michael@0 385 int xoffset,
michael@0 386 int yoffset,
michael@0 387 const unsigned char *dst_ptr,
michael@0 388 int dst_pixels_per_line,
michael@0 389 unsigned int *sse
michael@0 390 )
michael@0 391 {
michael@0 392 vp8_sub_pixel_variance16x16_wmt(src_ptr, src_pixels_per_line, xoffset, yoffset, dst_ptr, dst_pixels_per_line, sse);
michael@0 393 return *sse;
michael@0 394 }
michael@0 395
michael@0 396 unsigned int vp8_sub_pixel_variance16x8_wmt
michael@0 397 (
michael@0 398 const unsigned char *src_ptr,
michael@0 399 int src_pixels_per_line,
michael@0 400 int xoffset,
michael@0 401 int yoffset,
michael@0 402 const unsigned char *dst_ptr,
michael@0 403 int dst_pixels_per_line,
michael@0 404 unsigned int *sse
michael@0 405
michael@0 406 )
michael@0 407 {
michael@0 408 int xsum0, xsum1;
michael@0 409 unsigned int xxsum0, xxsum1;
michael@0 410
michael@0 411 if (xoffset == 4 && yoffset == 0)
michael@0 412 {
michael@0 413 vp8_half_horiz_variance16x_h_sse2(
michael@0 414 src_ptr, src_pixels_per_line,
michael@0 415 dst_ptr, dst_pixels_per_line, 8,
michael@0 416 &xsum0, &xxsum0);
michael@0 417 }
michael@0 418 else if (xoffset == 0 && yoffset == 4)
michael@0 419 {
michael@0 420 vp8_half_vert_variance16x_h_sse2(
michael@0 421 src_ptr, src_pixels_per_line,
michael@0 422 dst_ptr, dst_pixels_per_line, 8,
michael@0 423 &xsum0, &xxsum0);
michael@0 424 }
michael@0 425 else if (xoffset == 4 && yoffset == 4)
michael@0 426 {
michael@0 427 vp8_half_horiz_vert_variance16x_h_sse2(
michael@0 428 src_ptr, src_pixels_per_line,
michael@0 429 dst_ptr, dst_pixels_per_line, 8,
michael@0 430 &xsum0, &xxsum0);
michael@0 431 }
michael@0 432 else
michael@0 433 {
michael@0 434 vp8_filter_block2d_bil_var_sse2(
michael@0 435 src_ptr, src_pixels_per_line,
michael@0 436 dst_ptr, dst_pixels_per_line, 8,
michael@0 437 xoffset, yoffset,
michael@0 438 &xsum0, &xxsum0);
michael@0 439
michael@0 440 vp8_filter_block2d_bil_var_sse2(
michael@0 441 src_ptr + 8, src_pixels_per_line,
michael@0 442 dst_ptr + 8, dst_pixels_per_line, 8,
michael@0 443 xoffset, yoffset,
michael@0 444 &xsum1, &xxsum1);
michael@0 445 xsum0 += xsum1;
michael@0 446 xxsum0 += xxsum1;
michael@0 447 }
michael@0 448
michael@0 449 *sse = xxsum0;
michael@0 450 return (xxsum0 - (((unsigned int)xsum0 * xsum0) >> 7));
michael@0 451 }
michael@0 452
michael@0 453 unsigned int vp8_sub_pixel_variance8x16_wmt
michael@0 454 (
michael@0 455 const unsigned char *src_ptr,
michael@0 456 int src_pixels_per_line,
michael@0 457 int xoffset,
michael@0 458 int yoffset,
michael@0 459 const unsigned char *dst_ptr,
michael@0 460 int dst_pixels_per_line,
michael@0 461 unsigned int *sse
michael@0 462 )
michael@0 463 {
michael@0 464 int xsum;
michael@0 465 unsigned int xxsum;
michael@0 466
michael@0 467 if (xoffset == 4 && yoffset == 0)
michael@0 468 {
michael@0 469 vp8_half_horiz_variance8x_h_sse2(
michael@0 470 src_ptr, src_pixels_per_line,
michael@0 471 dst_ptr, dst_pixels_per_line, 16,
michael@0 472 &xsum, &xxsum);
michael@0 473 }
michael@0 474 else if (xoffset == 0 && yoffset == 4)
michael@0 475 {
michael@0 476 vp8_half_vert_variance8x_h_sse2(
michael@0 477 src_ptr, src_pixels_per_line,
michael@0 478 dst_ptr, dst_pixels_per_line, 16,
michael@0 479 &xsum, &xxsum);
michael@0 480 }
michael@0 481 else if (xoffset == 4 && yoffset == 4)
michael@0 482 {
michael@0 483 vp8_half_horiz_vert_variance8x_h_sse2(
michael@0 484 src_ptr, src_pixels_per_line,
michael@0 485 dst_ptr, dst_pixels_per_line, 16,
michael@0 486 &xsum, &xxsum);
michael@0 487 }
michael@0 488 else
michael@0 489 {
michael@0 490 vp8_filter_block2d_bil_var_sse2(
michael@0 491 src_ptr, src_pixels_per_line,
michael@0 492 dst_ptr, dst_pixels_per_line, 16,
michael@0 493 xoffset, yoffset,
michael@0 494 &xsum, &xxsum);
michael@0 495 }
michael@0 496
michael@0 497 *sse = xxsum;
michael@0 498 return (xxsum - (((unsigned int)xsum * xsum) >> 7));
michael@0 499 }
michael@0 500
michael@0 501
michael@0 502 unsigned int vp8_variance_halfpixvar16x16_h_wmt(
michael@0 503 const unsigned char *src_ptr,
michael@0 504 int src_pixels_per_line,
michael@0 505 const unsigned char *dst_ptr,
michael@0 506 int dst_pixels_per_line,
michael@0 507 unsigned int *sse)
michael@0 508 {
michael@0 509 int xsum0;
michael@0 510 unsigned int xxsum0;
michael@0 511
michael@0 512 vp8_half_horiz_variance16x_h_sse2(
michael@0 513 src_ptr, src_pixels_per_line,
michael@0 514 dst_ptr, dst_pixels_per_line, 16,
michael@0 515 &xsum0, &xxsum0);
michael@0 516
michael@0 517 *sse = xxsum0;
michael@0 518 return (xxsum0 - (((unsigned int)xsum0 * xsum0) >> 8));
michael@0 519 }
michael@0 520
michael@0 521
michael@0 522 unsigned int vp8_variance_halfpixvar16x16_v_wmt(
michael@0 523 const unsigned char *src_ptr,
michael@0 524 int src_pixels_per_line,
michael@0 525 const unsigned char *dst_ptr,
michael@0 526 int dst_pixels_per_line,
michael@0 527 unsigned int *sse)
michael@0 528 {
michael@0 529 int xsum0;
michael@0 530 unsigned int xxsum0;
michael@0 531 vp8_half_vert_variance16x_h_sse2(
michael@0 532 src_ptr, src_pixels_per_line,
michael@0 533 dst_ptr, dst_pixels_per_line, 16,
michael@0 534 &xsum0, &xxsum0);
michael@0 535
michael@0 536 *sse = xxsum0;
michael@0 537 return (xxsum0 - (((unsigned int)xsum0 * xsum0) >> 8));
michael@0 538 }
michael@0 539
michael@0 540
michael@0 541 unsigned int vp8_variance_halfpixvar16x16_hv_wmt(
michael@0 542 const unsigned char *src_ptr,
michael@0 543 int src_pixels_per_line,
michael@0 544 const unsigned char *dst_ptr,
michael@0 545 int dst_pixels_per_line,
michael@0 546 unsigned int *sse)
michael@0 547 {
michael@0 548 int xsum0;
michael@0 549 unsigned int xxsum0;
michael@0 550
michael@0 551 vp8_half_horiz_vert_variance16x_h_sse2(
michael@0 552 src_ptr, src_pixels_per_line,
michael@0 553 dst_ptr, dst_pixels_per_line, 16,
michael@0 554 &xsum0, &xxsum0);
michael@0 555
michael@0 556 *sse = xxsum0;
michael@0 557 return (xxsum0 - (((unsigned int)xsum0 * xsum0) >> 8));
michael@0 558 }

Mercurial Repository: The Tor Browser

Europalab SCM Repository Server

mercurial