hevcdsp: split the qpel functions by width instead of by the subpixel fraction
authorAnton Khirnov <anton@khirnov.net>
Fri, 24 Jul 2015 06:24:21 +0000 (08:24 +0200)
committerAnton Khirnov <anton@khirnov.net>
Sat, 5 Dec 2015 20:08:04 +0000 (21:08 +0100)
commit1f821750f0b8d0c87cbf88a28ad699b92db5ec88
tree78ec782cd9a422df28fa05c2c0a72650f1ec89c1
parent6788baebb3680d447eabdadf3f5743c8470a4611
hevcdsp: split the qpel functions by width instead of by the subpixel fraction

This should allow for more efficient SIMD.

Keep the C versions as they are now, to allow the compiler to inline the
interpolation coefficients.
libavcodec/hevc.c
libavcodec/hevcdsp.c
libavcodec/hevcdsp.h
libavcodec/hevcdsp_template.c