The DAV1D AV1 Video Decoder Has Become Very Fast

It was just two months ago that the VideoLAN/VLC crew announced the DAV1D AV1 video decoder and already it's becoming quite feature complete and super fast.

In a performance report shared today, DAV1D now covers the entire AV1 spec and all features for 8-bit and 10-bit depth.This decoder is up to 400% faster than the reference libaom decoder and usually at least 100% faster.

For reaching such insane performance, DAV1D implements a lot of hand-written Assembly code for modern desktop processors including AVX2. SSE/ARM optimizations are also being worked on DAV1D threading is also working for leveraging modern multi-core CPUs.

In terms of the state right now, "dav1d is very fast, dav1d is almost complete, dav1d is cool. We're finishing the rough edges for a release soon, so that we can hope that Firefox 65 will ship with dav1d for AV1 decoding. On the other platforms, SSE and ARM assembly will follow very quickly, and we're already as fast on ARMv8. Stay tuned for more!"

More details via this blog post. Yes, I'll be adding a DAV1D profile to the Phoronix Test Suite.