VideoLAN's dav1d 0.6 Released With More AVX2 + AVX-512 Optimizations
Dav1d already is quite fast on most CPUs but now is even faster on x86_64 hardware thanks to enabling AVX-512 optimizations and also extending their AVX2 and SSSE3 optimizations.
On the Arm front, Dav1d 0.6 has improved 10/12-bit decoding for ARMv8 hardware to the extent the performance can be up to 2.5x faster than the previous release.
More details on dav1d 0.6 at VideoLAN.org. I'll have up some fresh Dav1d 0.6 benchmarks shortly (the test profile is already updated).