diff options
| author | Kyle Siefring <kylesiefring@gmail.com> | 2017-10-07 16:02:02 -0400 |
|---|---|---|
| committer | Kyle Siefring <kylesiefring@gmail.com> | 2017-10-07 23:37:48 -0400 |
| commit | 9ca06bcdd21dc0e3c2baa80c4c7fd0e7e8637e59 (patch) | |
| tree | 29c5626341649a9b997c41926fd5c1679d269163 /vpx_dsp/vpx_dsp_rtcd_defs.pl | |
| parent | 807248ec81188ce12d7039e9b3c9d770e57fba5b (diff) | |
| download | libvpx-9ca06bcdd21dc0e3c2baa80c4c7fd0e7e8637e59.tar libvpx-9ca06bcdd21dc0e3c2baa80c4c7fd0e7e8637e59.tar.gz libvpx-9ca06bcdd21dc0e3c2baa80c4c7fd0e7e8637e59.tar.bz2 libvpx-9ca06bcdd21dc0e3c2baa80c4c7fd0e7e8637e59.zip | |
Add AVX2 version of vpx_convolve8_avg.
vpx_convolve8_avg works by first running a normal horizontal filter then a
vertical filter averages at the end.
The added vpx_convolve8_avg_avx2 calls pre-existing AVX2 code for the
horizontal step.
vpx_convolve8_avg_vert_avx2 is also added, but only uses ssse3 code.
Change-Id: If5160c0c8e778e10de61ee9bf42ee4be5975c983
Diffstat (limited to 'vpx_dsp/vpx_dsp_rtcd_defs.pl')
| -rw-r--r-- | vpx_dsp/vpx_dsp_rtcd_defs.pl | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/vpx_dsp/vpx_dsp_rtcd_defs.pl b/vpx_dsp/vpx_dsp_rtcd_defs.pl index 5cb17e167..43c506cd4 100644 --- a/vpx_dsp/vpx_dsp_rtcd_defs.pl +++ b/vpx_dsp/vpx_dsp_rtcd_defs.pl @@ -364,13 +364,13 @@ add_proto qw/void vpx_convolve8_vert/, "const uint8_t *src, ptrdiff_t src_stride specialize qw/vpx_convolve8_vert sse2 ssse3 avx2 neon dspr2 msa vsx/; add_proto qw/void vpx_convolve8_avg/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h"; -specialize qw/vpx_convolve8_avg sse2 ssse3 neon dspr2 msa vsx/; +specialize qw/vpx_convolve8_avg sse2 ssse3 avx2 neon dspr2 msa vsx/; add_proto qw/void vpx_convolve8_avg_horiz/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h"; specialize qw/vpx_convolve8_avg_horiz sse2 ssse3 neon dspr2 msa vsx/; add_proto qw/void vpx_convolve8_avg_vert/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h"; -specialize qw/vpx_convolve8_avg_vert sse2 ssse3 neon dspr2 msa vsx/; +specialize qw/vpx_convolve8_avg_vert sse2 ssse3 avx2 neon dspr2 msa vsx/; add_proto qw/void vpx_scaled_2d/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h"; specialize qw/vpx_scaled_2d ssse3 neon/; |