shithub: libvpx

Info  •  Files  •  Log  •  Branches

Changes 4-tap SSSE3 filter to 8-tap AVX2 filter.

AVX2's 8-tap filter is slightly faster than 4-tap SSSE3 filter.

Change-Id: I5fc37c431670780108706b206b32c791828555c9

--- a/vpx_dsp/x86/vpx_subpixel_8t_intrin_avx2.c

+++ b/vpx_dsp/x86/vpx_subpixel_8t_intrin_avx2.c

@@ -376,20 +376,7 @@

 #define vpx_filter_block1d8_h2_avg_avx2 vpx_filter_block1d8_h2_avg_ssse3

 #define vpx_filter_block1d4_v2_avg_avx2 vpx_filter_block1d4_v2_avg_ssse3

 #define vpx_filter_block1d4_h2_avg_avx2 vpx_filter_block1d4_h2_avg_ssse3

-#if HAVE_SSSE3

-filter8_1dfunction vpx_filter_block1d16_v4_ssse3;

-filter8_1dfunction vpx_filter_block1d16_h4_ssse3;

-filter8_1dfunction vpx_filter_block1d8_v4_ssse3;

-filter8_1dfunction vpx_filter_block1d8_h4_ssse3;

-filter8_1dfunction vpx_filter_block1d4_v4_ssse3;

-filter8_1dfunction vpx_filter_block1d4_h4_ssse3;

-#define vpx_filter_block1d16_v4_avx2 vpx_filter_block1d16_v4_ssse3

-#define vpx_filter_block1d16_h4_avx2 vpx_filter_block1d16_h4_ssse3

-#define vpx_filter_block1d8_v4_avx2 vpx_filter_block1d8_v4_ssse3

-#define vpx_filter_block1d8_h4_avx2 vpx_filter_block1d8_h4_ssse3

-#define vpx_filter_block1d4_v4_avx2 vpx_filter_block1d4_v4_ssse3

-#define vpx_filter_block1d4_h4_avx2 vpx_filter_block1d4_h4_ssse3

-#else

+

 #define vpx_filter_block1d16_v4_avx2 vpx_filter_block1d16_v8_avx2

 #define vpx_filter_block1d16_h4_avx2 vpx_filter_block1d16_h8_avx2

 #define vpx_filter_block1d8_v4_avx2 vpx_filter_block1d8_v8_avx2

@@ -396,7 +383,6 @@

 #define vpx_filter_block1d8_h4_avx2 vpx_filter_block1d8_h8_avx2

 #define vpx_filter_block1d4_v4_avx2 vpx_filter_block1d4_v8_avx2

 #define vpx_filter_block1d4_h4_avx2 vpx_filter_block1d4_h8_avx2

-#endif

 #define vpx_filter_block1d16_v4_avg_avx2 vpx_filter_block1d16_v8_avg_avx2

 #define vpx_filter_block1d16_h4_avg_avx2 vpx_filter_block1d16_h8_avg_avx2

 #define vpx_filter_block1d8_v4_avg_avx2 vpx_filter_block1d8_v8_avg_avx2