You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

107 lines
3.2KB

  1. /*
  2. * Copyright (c) 2016 Alexandra Hájková
  3. *
  4. * This file is part of FFmpeg.
  5. *
  6. * FFmpeg is free software; you can redistribute it and/or modify
  7. * it under the terms of the GNU General Public License as published by
  8. * the Free Software Foundation; either version 2 of the License, or
  9. * (at your option) any later version.
  10. *
  11. * FFmpeg is distributed in the hope that it will be useful,
  12. * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  14. * GNU General Public License for more details.
  15. *
  16. * You should have received a copy of the GNU General Public License along
  17. * with FFmpeg; if not, write to the Free Software Foundation, Inc.,
  18. * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
  19. */
  20. #include <string.h>
  21. #include "libavutil/intreadwrite.h"
  22. #include "libavutil/mem_internal.h"
  23. #include "libavcodec/hevcdsp.h"
  24. #include "checkasm.h"
  25. #define randomize_buffers(buf, size) \
  26. do { \
  27. int j; \
  28. for (j = 0; j < size; j++) { \
  29. int16_t r = rnd(); \
  30. AV_WN16A(buf + j, r); \
  31. } \
  32. } while (0)
  33. static void check_idct(HEVCDSPContext h, int bit_depth)
  34. {
  35. int i;
  36. LOCAL_ALIGNED(32, int16_t, coeffs0, [32 * 32]);
  37. LOCAL_ALIGNED(32, int16_t, coeffs1, [32 * 32]);
  38. for (i = 2; i <= 5; i++) {
  39. int block_size = 1 << i;
  40. int size = block_size * block_size;
  41. int col_limit = block_size;
  42. declare_func(void, int16_t *coeffs, int col_limit);
  43. randomize_buffers(coeffs0, size);
  44. memcpy(coeffs1, coeffs0, sizeof(*coeffs0) * size);
  45. if (check_func(h.idct[i - 2], "hevc_idct_%dx%d_%d", block_size, block_size, bit_depth)) {
  46. call_ref(coeffs0, col_limit);
  47. call_new(coeffs1, col_limit);
  48. if (memcmp(coeffs0, coeffs1, sizeof(*coeffs0) * size))
  49. fail();
  50. bench_new(coeffs1, col_limit);
  51. }
  52. }
  53. }
  54. static void check_idct_dc(HEVCDSPContext h, int bit_depth)
  55. {
  56. int i;
  57. LOCAL_ALIGNED(32, int16_t, coeffs0, [32 * 32]);
  58. LOCAL_ALIGNED(32, int16_t, coeffs1, [32 * 32]);
  59. for (i = 2; i <= 5; i++) {
  60. int block_size = 1 << i;
  61. int size = block_size * block_size;
  62. declare_func_emms(AV_CPU_FLAG_MMXEXT, void, int16_t *coeffs);
  63. randomize_buffers(coeffs0, size);
  64. memcpy(coeffs1, coeffs0, sizeof(*coeffs0) * size);
  65. if (check_func(h.idct_dc[i - 2], "hevc_idct_%dx%d_dc_%d", block_size, block_size, bit_depth)) {
  66. call_ref(coeffs0);
  67. call_new(coeffs1);
  68. if (memcmp(coeffs0, coeffs1, sizeof(*coeffs0) * size))
  69. fail();
  70. bench_new(coeffs1);
  71. }
  72. }
  73. }
  74. void checkasm_check_hevc_idct(void)
  75. {
  76. int bit_depth;
  77. for (bit_depth = 8; bit_depth <= 12; bit_depth += 2) {
  78. HEVCDSPContext h;
  79. ff_hevc_dsp_init(&h, bit_depth);
  80. check_idct_dc(h, bit_depth);
  81. }
  82. report("idct_dc");
  83. for (bit_depth = 8; bit_depth <= 12; bit_depth += 2) {
  84. HEVCDSPContext h;
  85. ff_hevc_dsp_init(&h, bit_depth);
  86. check_idct(h, bit_depth);
  87. }
  88. report("idct");
  89. }