diff options
| author | Jingning Han <jingning@google.com> | 2017-02-01 12:19:01 -0800 |
|---|---|---|
| committer | Jingning Han <jingning@google.com> | 2017-02-01 14:47:53 -0800 |
| commit | 8f953897429f75ce9f44e80feaae810a6c9752bd (patch) | |
| tree | ec6cfc17323f3bc588d147ec1637870b135e4857 /vpx_dsp/vpx_dsp_rtcd_defs.pl | |
| parent | a7949f2dd2bb68516a7ab4bc8c60f9d72231b800 (diff) | |
| download | libvpx-8f953897429f75ce9f44e80feaae810a6c9752bd.tar libvpx-8f953897429f75ce9f44e80feaae810a6c9752bd.tar.gz libvpx-8f953897429f75ce9f44e80feaae810a6c9752bd.tar.bz2 libvpx-8f953897429f75ce9f44e80feaae810a6c9752bd.zip | |
Add SSSE3 intrinsic 8x8 inverse 2D-DCT
The intrinsic version reduces the average cycles from 183 to 175.
Change-Id: I7c1bcdb0a830266e93d8347aed38120fb3be0e03
Diffstat (limited to 'vpx_dsp/vpx_dsp_rtcd_defs.pl')
| -rw-r--r-- | vpx_dsp/vpx_dsp_rtcd_defs.pl | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/vpx_dsp/vpx_dsp_rtcd_defs.pl b/vpx_dsp/vpx_dsp_rtcd_defs.pl index 3cb2011b8..a5b79ba4e 100644 --- a/vpx_dsp/vpx_dsp_rtcd_defs.pl +++ b/vpx_dsp/vpx_dsp_rtcd_defs.pl @@ -678,8 +678,9 @@ if (vpx_config("CONFIG_VP9_HIGHBITDEPTH") eq "yes") { add_proto qw/void vpx_idct4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int stride"; specialize qw/vpx_idct4x4_1_add neon sse2/; + # TODO(jingning): Add ssse3 for high bit-depth add_proto qw/void vpx_idct8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int stride"; - specialize qw/vpx_idct8x8_64_add neon sse2/, "$ssse3_x86_64"; + specialize qw/vpx_idct8x8_64_add neon sse2/; add_proto qw/void vpx_idct8x8_12_add/, "const tran_low_t *input, uint8_t *dest, int stride"; specialize qw/vpx_idct8x8_12_add neon sse2/, "$ssse3_x86_64"; @@ -766,7 +767,7 @@ if (vpx_config("CONFIG_VP9_HIGHBITDEPTH") eq "yes") { specialize qw/vpx_idct8x8_1_add sse2 neon dspr2 msa/; add_proto qw/void vpx_idct8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int stride"; - specialize qw/vpx_idct8x8_64_add sse2 neon dspr2 msa/, "$ssse3_x86_64"; + specialize qw/vpx_idct8x8_64_add sse2 ssse3 neon dspr2 msa/; add_proto qw/void vpx_idct8x8_12_add/, "const tran_low_t *input, uint8_t *dest, int stride"; specialize qw/vpx_idct8x8_12_add sse2 neon dspr2 msa/, "$ssse3_x86_64"; |