Hacker News Reader

And MoltenVK isn't just an experiment or hobby project, it's already being used in production—one prominent example is Autodesk Fusion 360, which even uses it on M1.

And, Kompute (kompute.cc) is a portable substitute for CUDA, given Vulkan 1.1.

I wonder now whether the Vulkan API is an adequate interface to securely expose GPU capabilities to multiple VMs, i.e. without those VMs being able to see other VMs' memory. If not, what would need to happen to get there?

Apple is holding back both the open web and open graphics. They're a cancer on the industry (and why I'll never give them a penny nor ever develop for one of their platforms).