
Merged this week were several ffv1enc_vulkan commits to clean-up the code and make various minor optimizations and other improvements.
Most notable for end-users though is this commit to the FFV1 encoder to make use of the cached EC writer from the decoder. As noted in that commit:
"ffv1enc_vulkan: implement the cached EC writer from the decoder
This gives a 35% speedup on AMD and 50% on Nvidia."
It's quite a win with a 35% improvement on AMD GPus and 50% for NVIDIA GPUs. No word on Intel performance benefits from this Vulkan encode optimization.