h264qpel.c 3.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081
  1. /*
  2. * Copyright (c) 2015 Henrik Gramner
  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 "checkasm.h"
  22. #include "libavcodec/h264qpel.h"
  23. #include "libavutil/common.h"
  24. #include "libavutil/internal.h"
  25. #include "libavutil/intreadwrite.h"
  26. static const uint32_t pixel_mask[3] = { 0xffffffff, 0x01ff01ff, 0x03ff03ff };
  27. #define SIZEOF_PIXEL ((bit_depth + 7) / 8)
  28. #define BUF_SIZE (2 * 16 * (16 + 3 + 4))
  29. #define randomize_buffers() \
  30. do { \
  31. uint32_t mask = pixel_mask[bit_depth - 8]; \
  32. int k; \
  33. for (k = 0; k < BUF_SIZE; k += 4) { \
  34. uint32_t r = rnd() & mask; \
  35. AV_WN32A(buf0 + k, r); \
  36. AV_WN32A(buf1 + k, r); \
  37. r = rnd(); \
  38. AV_WN32A(dst0 + k, r); \
  39. AV_WN32A(dst1 + k, r); \
  40. } \
  41. } while (0)
  42. #define src0 (buf0 + 3 * 2 * 16) /* h264qpel functions read data from negative src pointer offsets */
  43. #define src1 (buf1 + 3 * 2 * 16)
  44. void checkasm_check_h264qpel(void)
  45. {
  46. LOCAL_ALIGNED_16(uint8_t, buf0, [BUF_SIZE]);
  47. LOCAL_ALIGNED_16(uint8_t, buf1, [BUF_SIZE]);
  48. LOCAL_ALIGNED_16(uint8_t, dst0, [BUF_SIZE]);
  49. LOCAL_ALIGNED_16(uint8_t, dst1, [BUF_SIZE]);
  50. H264QpelContext h;
  51. int op, bit_depth, i, j;
  52. declare_func_emms(AV_CPU_FLAG_MMX | AV_CPU_FLAG_MMXEXT, void, uint8_t *dst, const uint8_t *src, ptrdiff_t stride);
  53. for (op = 0; op < 2; op++) {
  54. qpel_mc_func (*tab)[16] = op ? h.avg_h264_qpel_pixels_tab : h.put_h264_qpel_pixels_tab;
  55. const char *op_name = op ? "avg" : "put";
  56. for (bit_depth = 8; bit_depth <= 10; bit_depth++) {
  57. ff_h264qpel_init(&h, bit_depth);
  58. for (i = 0; i < (op ? 3 : 4); i++) {
  59. int size = 16 >> i;
  60. for (j = 0; j < 16; j++)
  61. if (check_func(tab[i][j], "%s_h264_qpel_%d_mc%d%d_%d", op_name, size, j & 3, j >> 2, bit_depth)) {
  62. randomize_buffers();
  63. call_ref(dst0, src0, size * SIZEOF_PIXEL);
  64. call_new(dst1, src1, size * SIZEOF_PIXEL);
  65. if (memcmp(buf0, buf1, BUF_SIZE) || memcmp(dst0, dst1, BUF_SIZE))
  66. fail();
  67. bench_new(dst1, src1, size * SIZEOF_PIXEL);
  68. }
  69. }
  70. }
  71. report("%s", op_name);
  72. }
  73. }