Hacker News Reader

We are not better off with CUDA working in My WSL. That keeps us confined to NVIDIA.

It suffices to have full Vulkan support. Then we can use Kompute, which is portable to any GPU, not just NVIDIA.

From the tutorial it is unclear what Kompute adds over raw Vulkan apart from Tensor type.

Its value is not so much in what it adds, but in what it subtracts: 2000 lines of boilerplate.

Why Kompute and not OpenCL ?

Powerful forces drive Vulkan implementation. OpenCL is always an afterthought, where it even happens at all.