dav1d 0.8.0 'Eurasian hobby' the fast and lean AV1 decoder

This is a major update of the dav1d, the fast and lean AV1 decoder,
codename 'Eurasian hobby'.

This release improves, once again, the speed on all platforms.

On the desktop platforms, it improves the performance (up to 10%)
by doing less allocations and adding more optimizations for AVX2 and SSE.

On the ARM platforms, this release brings numerous improvements for ARM32
and notably optimizes most of the 10/12bit bitdepth filters.

dav1d 0.8.0 'Eurasian hobby' the fast and lean AV1 decoder

This is a major update of the dav1d, the fast and lean AV1 decoder, codename 'Eurasian hobby'.

0.8.0 improves the performance by using a picture buffer pool. Up to 10% faster on Windows and less on other platforms.

It also adds support for Apple ARM Silicon and extends optimizations on ARM32 (8-bit ipred paeth, smooth, cfl, and 10/12-bit mc_avg/mask/w_avg, put/prep 8tap/bilin, wiener and CDEF filters), ARM64 (cfl_ac 444 for all bitdepths) and x86 (MC 8-tap, mc_scaled in AVX2, CDEF in SSE and {put/prep}_{8tap/bilin} in SSSE3).