Is that due to some kind of issue with the architecture, or just a matter of software support?
In the latter case, I'd expect patches for AMD or Intel to become a priority pretty quickly. After all, they need their products to run on systems that customers can buy.
Presumably you mean Intel’s integrated GPUs? They do have the Arc line of discrete GPUs now, and those are a bit more than a frame buffer with a clever blitter.
Is that due to some kind of issue with the architecture, or just a matter of software support?
In the latter case, I'd expect patches for AMD or Intel to become a priority pretty quickly. After all, they need their products to run on systems that customers can buy.
Well, they don't support CUDA and I don't see CUDA coming to AMD any time soon.
Intel is just plain not capable of it because it's not really a GPU, more a framebuffer with a clever blitter.
Presumably you mean Intel’s integrated GPUs? They do have the Arc line of discrete GPUs now, and those are a bit more than a frame buffer with a clever blitter.
Great opportunity to adopt Khronos APIs.
Isn't OpenCL is a widely supported compute platform alternative to CUDA? Not sure if it's fair comparison.
ZLUDA?