diff options
| author | Jingning Han <jingning@google.com> | 2014-04-23 09:13:45 -0700 |
|---|---|---|
| committer | Jingning Han <jingning@google.com> | 2014-04-29 15:49:18 -0700 |
| commit | 1eaa3a76dc50fc8451cedbf4797d7e3b7a2d77d7 (patch) | |
| tree | c019122dae2c4d611bb6a882b594fb5634bf59f3 /vp9/common | |
| parent | e38ca5422cebb7411f9bb501ce21f63f1358e5dc (diff) | |
| download | libvpx-1eaa3a76dc50fc8451cedbf4797d7e3b7a2d77d7.tar libvpx-1eaa3a76dc50fc8451cedbf4797d7e3b7a2d77d7.tar.gz libvpx-1eaa3a76dc50fc8451cedbf4797d7e3b7a2d77d7.tar.bz2 libvpx-1eaa3a76dc50fc8451cedbf4797d7e3b7a2d77d7.zip | |
Enable SSSE3 implementation of 8x8 forward 2D-DCT
Assembly implementation of ssse3 8x8 forward 2D-DCT. The current
version is turned on only for x86_64. The average unit runtime
goes from 157 cycles down to 136 cycles, i.e., about 12.8% faster.
This translates into about 1.5% speed-up for pedestrian_area 1080p
at speed 2.
Change-Id: I0f12435857e9425ed7ce12541344dfa16837f4f4
Diffstat (limited to 'vp9/common')
| -rw-r--r-- | vp9/common/vp9_rtcd_defs.pl | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/vp9/common/vp9_rtcd_defs.pl b/vp9/common/vp9_rtcd_defs.pl index 8a8155410..0b8a025a0 100644 --- a/vp9/common/vp9_rtcd_defs.pl +++ b/vp9/common/vp9_rtcd_defs.pl @@ -743,7 +743,7 @@ add_proto qw/void vp9_fdct4x4/, "const int16_t *input, int16_t *output, int stri specialize qw/vp9_fdct4x4 sse2 avx2/; add_proto qw/void vp9_fdct8x8/, "const int16_t *input, int16_t *output, int stride"; -specialize qw/vp9_fdct8x8 sse2 avx2/; +specialize qw/vp9_fdct8x8 sse2 avx2/, "$ssse3_x86_64"; add_proto qw/void vp9_fdct16x16/, "const int16_t *input, int16_t *output, int stride"; specialize qw/vp9_fdct16x16 sse2 avx2/; |