Alternative(s) to run CUDA on non-Nvidia hardware

woctordho 22 minutes ago | parent | next [-]

There's nothing wrong to run CUDA on non-Nvidia hardware. CUDA has an interface that is reasonably well-designed, well-documented/reverse-engineered, and battle-tested for decades. What we need is not to invent another interface just under the name of 'open standard', but to implement the same interface. ROCm is exactly doing this, and so are other hardware SDKs such as MooreThread and Alibaba T-Head.

▲

luciana1u 22 minutes ago | parent | prev | next [-]

every CUDA alternative follows the same arc: bold launch, works for 3 operations, then a Discord server where the last message is 'any updates?' from 2024

▲

pjmlp an hour ago | parent | prev | next [-]

Most of these "alternatives" focus on CUDA C++, and overlook what actually makes CUDA interesting.

Already in 2020,

https://developer.nvidia.com/blog/cuda-refresher-the-gpu-com...

▲

msond an hour ago | parent [-]

We're actually targeting all of it, and not just CUDA C++.

▲

pjmlp an hour ago | parent | next [-]

Including stuff like Fortran, Haskell, Java, .NET via PTX, Python JIT, IDE tooling integration with major IDEs, graphical GPU debugging and profiling, libraries and co?

Then I guess all the best.

▲

zorked 28 minutes ago | parent [-]

This post has some serious peanut-gallery vibes.

	▲	pjmlp 8 minutes ago \| parent [-]
		Peanut-gallery is happily using CUDA, and needs actual sound reasons to move.

▲

embedding-shape an hour ago | parent | prev [-]

Ambitions but neat, good luck if nothing else :)

If you were to guess, when do you think your Nsight Compute alternative might be ready with your own toolchain?

▲

maxloh 40 minutes ago | parent | prev | next [-]

There is also ZLUDA, which is open source and works on pre-compiled binaries.

https://github.com/vosen/ZLUDA

▲

DiabloD3 29 minutes ago | parent | prev | next [-]

Its easier to just get rid of your legacy code entirely and use Vulkan for compute, or have your compiler emit SPIR-V directly.

No reason to tie yourself to Nvidia's moat.

	▲	swerner 23 minutes ago \| parent [-]
		Unfortunately, Vulkan Compute doesn’t to all the things that OpenCL, SYCL, HIP or CUDA do.

▲

lulzx 28 minutes ago | parent | prev [-]

I have been trying for cuda -> metal, to run it on mac, https://github.com/lulzx/cuda-metal