Last released Oct 7, 2022
Use Vulkan to allocate CUDA memory that is writable from the host
Supported by