| OLD | NEW |
| 1 /* | 1 /* |
| 2 * Copyright (c) 2013 The WebM project authors. All Rights Reserved. | 2 * Copyright (c) 2013 The WebM project authors. All Rights Reserved. |
| 3 * | 3 * |
| 4 * Use of this source code is governed by a BSD-style license | 4 * Use of this source code is governed by a BSD-style license |
| 5 * that can be found in the LICENSE file in the root of the source | 5 * that can be found in the LICENSE file in the root of the source |
| 6 * tree. An additional intellectual property rights grant can be found | 6 * tree. An additional intellectual property rights grant can be found |
| 7 * in the file PATENTS. All contributing project authors may | 7 * in the file PATENTS. All contributing project authors may |
| 8 * be found in the AUTHORS file in the root of the source tree. | 8 * be found in the AUTHORS file in the root of the source tree. |
| 9 */ | 9 */ |
| 10 | 10 |
| 11 #include <assert.h> | 11 #include <assert.h> |
| 12 #include <stdio.h> | 12 #include <stdio.h> |
| 13 | 13 |
| 14 #include "./vpx_dsp_rtcd.h" | 14 #include "./vpx_dsp_rtcd.h" |
| 15 #include "vpx_dsp/mips/vpx_common_dspr2.h" | 15 #include "vpx_dsp/mips/convolve_common_dspr2.h" |
| 16 #include "vpx_dsp/vpx_dsp_common.h" | 16 #include "vpx_dsp/vpx_dsp_common.h" |
| 17 #include "vpx_dsp/vpx_filter.h" | 17 #include "vpx_dsp/vpx_filter.h" |
| 18 #include "vpx_ports/mem.h" | 18 #include "vpx_ports/mem.h" |
| 19 | 19 |
| 20 #if HAVE_DSPR2 | 20 #if HAVE_DSPR2 |
| 21 static void convolve_vert_4_dspr2(const uint8_t *src, | 21 static void convolve_vert_4_dspr2(const uint8_t *src, |
| 22 int32_t src_stride, | 22 int32_t src_stride, |
| 23 uint8_t *dst, | 23 uint8_t *dst, |
| 24 int32_t dst_stride, | 24 int32_t dst_stride, |
| 25 const int16_t *filter_y, | 25 const int16_t *filter_y, |
| (...skipping 300 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
| 326 src += src_stride; | 326 src += src_stride; |
| 327 dst += dst_stride; | 327 dst += dst_stride; |
| 328 } | 328 } |
| 329 } | 329 } |
| 330 | 330 |
| 331 void vpx_convolve8_vert_dspr2(const uint8_t *src, ptrdiff_t src_stride, | 331 void vpx_convolve8_vert_dspr2(const uint8_t *src, ptrdiff_t src_stride, |
| 332 uint8_t *dst, ptrdiff_t dst_stride, | 332 uint8_t *dst, ptrdiff_t dst_stride, |
| 333 const int16_t *filter_x, int x_step_q4, | 333 const int16_t *filter_x, int x_step_q4, |
| 334 const int16_t *filter_y, int y_step_q4, | 334 const int16_t *filter_y, int y_step_q4, |
| 335 int w, int h) { | 335 int w, int h) { |
| 336 if (((const int32_t *)filter_y)[1] == 0x800000) { | 336 assert(y_step_q4 == 16); |
| 337 vpx_convolve_copy(src, src_stride, | 337 assert(((const int32_t *)filter_y)[1] != 0x800000); |
| 338 dst, dst_stride, | 338 |
| 339 filter_x, x_step_q4, | 339 if (((const int32_t *)filter_y)[0] == 0) { |
| 340 filter_y, y_step_q4, | |
| 341 w, h); | |
| 342 } else if (((const int32_t *)filter_y)[0] == 0) { | |
| 343 vpx_convolve2_vert_dspr2(src, src_stride, | 340 vpx_convolve2_vert_dspr2(src, src_stride, |
| 344 dst, dst_stride, | 341 dst, dst_stride, |
| 345 filter_x, x_step_q4, | 342 filter_x, x_step_q4, |
| 346 filter_y, y_step_q4, | 343 filter_y, y_step_q4, |
| 347 w, h); | 344 w, h); |
| 348 } else { | 345 } else { |
| 349 if (16 == y_step_q4) { | 346 uint32_t pos = 38; |
| 350 uint32_t pos = 38; | |
| 351 | 347 |
| 352 /* bit positon for extract from acc */ | 348 /* bit positon for extract from acc */ |
| 353 __asm__ __volatile__ ( | 349 __asm__ __volatile__ ( |
| 354 "wrdsp %[pos], 1 \n\t" | 350 "wrdsp %[pos], 1 \n\t" |
| 355 : | 351 : |
| 356 : [pos] "r" (pos) | 352 : [pos] "r" (pos) |
| 357 ); | 353 ); |
| 358 | 354 |
| 359 prefetch_store(dst); | 355 prefetch_store(dst); |
| 360 | 356 |
| 361 switch (w) { | 357 switch (w) { |
| 362 case 4 : | 358 case 4 : |
| 363 case 8 : | 359 case 8 : |
| 364 case 16 : | 360 case 16 : |
| 365 case 32 : | 361 case 32 : |
| 366 convolve_vert_4_dspr2(src, src_stride, | 362 convolve_vert_4_dspr2(src, src_stride, |
| 367 dst, dst_stride, | 363 dst, dst_stride, |
| 368 filter_y, w, h); | 364 filter_y, w, h); |
| 369 break; | 365 break; |
| 370 case 64 : | 366 case 64 : |
| 371 prefetch_store(dst + 32); | 367 prefetch_store(dst + 32); |
| 372 convolve_vert_64_dspr2(src, src_stride, | 368 convolve_vert_64_dspr2(src, src_stride, |
| 373 dst, dst_stride, | |
| 374 filter_y, h); | |
| 375 break; | |
| 376 default: | |
| 377 vpx_convolve8_vert_c(src, src_stride, | |
| 378 dst, dst_stride, | 369 dst, dst_stride, |
| 379 filter_x, x_step_q4, | 370 filter_y, h); |
| 380 filter_y, y_step_q4, | 371 break; |
| 381 w, h); | 372 default: |
| 382 break; | 373 vpx_convolve8_vert_c(src, src_stride, |
| 383 } | 374 dst, dst_stride, |
| 384 } else { | 375 filter_x, x_step_q4, |
| 385 vpx_convolve8_vert_c(src, src_stride, | 376 filter_y, y_step_q4, |
| 386 dst, dst_stride, | 377 w, h); |
| 387 filter_x, x_step_q4, | 378 break; |
| 388 filter_y, y_step_q4, | |
| 389 w, h); | |
| 390 } | 379 } |
| 391 } | 380 } |
| 392 } | 381 } |
| 393 | 382 |
| 394 #endif | 383 #endif |
| OLD | NEW |