mozilla
/
aom
зеркало из
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

mips msa vp9 avg subpel variance optimization rebased 

Change-Id: Ia21987010dbb688e2a8fa204ca9129d2f34c9581
This commit is contained in:
Parag Salasakar 2015-07-08 12:07:28 +05:30
Родитель 892128f6ca
Коммит 1d7f1ca7da
3 изменённых файлов: 1291 добавлений и 13 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

42
test/variance_test.cc
Просмотреть файл

@ -1994,5 +1994,47 @@ INSTANTIATE_TEST_CASE_P(
make_tuple(5, 6, subpel_variance32x64_msa, 0), make_tuple(5, 6, subpel_variance32x64_msa, 0),
make_tuple(6, 5, subpel_variance64x32_msa, 0), make_tuple(6, 5, subpel_variance64x32_msa, 0),
make_tuple(6, 6, subpel_variance64x64_msa, 0))); make_tuple(6, 6, subpel_variance64x64_msa, 0)));
const SubpixAvgVarMxNFunc subpel_avg_variance64x64_msa =
vpx_sub_pixel_avg_variance64x64_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance64x32_msa =
vpx_sub_pixel_avg_variance64x32_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance32x64_msa =
vpx_sub_pixel_avg_variance32x64_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance32x32_msa =
vpx_sub_pixel_avg_variance32x32_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance32x16_msa =
vpx_sub_pixel_avg_variance32x16_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance16x32_msa =
vpx_sub_pixel_avg_variance16x32_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance16x16_msa =
vpx_sub_pixel_avg_variance16x16_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance16x8_msa =
vpx_sub_pixel_avg_variance16x8_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance8x16_msa =
vpx_sub_pixel_avg_variance8x16_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance8x8_msa =
vpx_sub_pixel_avg_variance8x8_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance8x4_msa =
vpx_sub_pixel_avg_variance8x4_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance4x8_msa =
vpx_sub_pixel_avg_variance4x8_msa;
const SubpixAvgVarMxNFunc subpel_avg_variance4x4_msa =
vpx_sub_pixel_avg_variance4x4_msa;
INSTANTIATE_TEST_CASE_P(
MSA, VpxSubpelAvgVarianceTest,
::testing::Values(make_tuple(6, 6, subpel_avg_variance64x64_msa, 0),
make_tuple(6, 5, subpel_avg_variance64x32_msa, 0),
make_tuple(5, 6, subpel_avg_variance32x64_msa, 0),
make_tuple(5, 5, subpel_avg_variance32x32_msa, 0),
make_tuple(5, 4, subpel_avg_variance32x16_msa, 0),
make_tuple(4, 5, subpel_avg_variance16x32_msa, 0),
make_tuple(4, 4, subpel_avg_variance16x16_msa, 0),
make_tuple(4, 3, subpel_avg_variance16x8_msa, 0),
make_tuple(3, 4, subpel_avg_variance8x16_msa, 0),
make_tuple(3, 3, subpel_avg_variance8x8_msa, 0),
make_tuple(3, 2, subpel_avg_variance8x4_msa, 0),
make_tuple(2, 3, subpel_avg_variance4x8_msa, 0),
make_tuple(2, 2, subpel_avg_variance4x4_msa, 0)));
#endif // HAVE_MSA #endif // HAVE_MSA
} // namespace } // namespace

1236
vpx_dsp/mips/sub_pixel_variance_msa.c
Просмотреть файл

Разница между файлами не показана из-за своего большого размера Загрузить разницу

26
vpx_dsp/vpx_dsp_rtcd_defs.pl
Просмотреть файл

@ -526,43 +526,43 @@ add_proto qw/uint32_t vpx_sub_pixel_variance4x4/, "const uint8_t *src_ptr, int s
specialize qw/vpx_sub_pixel_variance4x4 mmx msa/, "$sse_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_variance4x4 mmx msa/, "$sse_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance64x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance64x64 avx2/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance64x64 avx2 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance64x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance64x32/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance64x32 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance32x64/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance32x64/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance32x64 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance32x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance32x32 avx2/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance32x32 avx2 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance32x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance32x16/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance32x16 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance16x32/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance16x32/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance16x32 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance16x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance16x16/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance16x16 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance16x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance16x8/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance16x8 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance8x16/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance8x16/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance8x16 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance8x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance8x8/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance8x8 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance8x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance8x4/, "$sse2_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance8x4 msa/, "$sse2_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance4x8/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance4x8/, "$sse_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance4x8 msa/, "$sse_x86inc", "$ssse3_x86inc";
add_proto qw/uint32_t vpx_sub_pixel_avg_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred"; add_proto qw/uint32_t vpx_sub_pixel_avg_variance4x4/, "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, uint32_t *sse, const uint8_t *second_pred";
specialize qw/vpx_sub_pixel_avg_variance4x4/, "$sse_x86inc", "$ssse3_x86inc"; specialize qw/vpx_sub_pixel_avg_variance4x4 msa/, "$sse_x86inc", "$ssse3_x86inc";
# #
# Specialty Subpixel # Specialty Subpixel