diff options
| author | David Barker <david.barker@argondesign.com> | 2017-05-18 15:12:50 +0100 |
|---|---|---|
| committer | Debargha Mukherjee <debargha@google.com> | 2017-05-23 13:22:36 +0000 |
| commit | 5d34e6a738984c6efd3884d7703680687cfdc00c (patch) | |
| tree | fa5de29dad405b30a42c0e10f0f04693b8907718 /aom_dsp/aom_convolve.c | |
| parent | 0f3c94e134ab4d358be43c39f2fed82fb8cdcdf5 (diff) | |
Vectorize high-precision convolve filter
Add SSE2 lowbd and SSSE3 highbd versions of the filters
introduced in https://aomedia-review.googlesource.com/c/11962/ .
These filters are equivalent in speed to the SSE2 implementations
of the regular convolve filter. The average time to filter a
64x64 block is:
lowbd C: 52us
lowbd SSE2: 5.6us
highbd C: 53us
highbd SSSE3: 5.8us
Also add a correctness test based on the warp filter tests.
Change-Id: Ia0d81100e8a414bbfc2b5f664d751cf24765299e
Diffstat (limited to 'aom_dsp/aom_convolve.c')
| -rw-r--r-- | aom_dsp/aom_convolve.c | 2 |
1 files changed, 0 insertions, 2 deletions
diff --git a/aom_dsp/aom_convolve.c b/aom_dsp/aom_convolve.c index 7182d7b3a..1abd9a241 100644 --- a/aom_dsp/aom_convolve.c +++ b/aom_dsp/aom_convolve.c @@ -447,8 +447,6 @@ void aom_convolve8_add_src_c(const uint8_t *src, ptrdiff_t src_stride, x_step_q4, filters_y, y0_q4, y_step_q4, w, h); } -#define EXTRAPREC_BITS 2 -#define EXTRAPREC_CLAMP_LIMIT (512 << EXTRAPREC_BITS) static void convolve_add_src_horiz_hip(const uint8_t *src, ptrdiff_t src_stride, uint16_t *dst, ptrdiff_t dst_stride, const InterpKernel *x_filters, int x0_q4, |