Live data from Hacker News

Zluda 6 release (run unmodified CUDA applications on non-Nvidia GPUs)

vosen.github.io

11–16 of 16 posts

Re: Zluda 6 release (run unmodified CUDA applications on non-Nvidia GPUs)

#12

Isn't this tool directly violating NVIDIA license terms? I thought they disallow running CUDA on non-NVIDIA hardware. Or is it just because there's no money for NVIDIA in suing poor, Eastern European programmer?

ZLUDA was funded by AMD for several years. Somewhat ironically, AMD's legal team went after them for releasing that code as open source and the developer rebuilt it from a pre-AMD funded version.

As to whether it's violating the EULA terms. Maybe, depends exactly what they did and how they did it. The better question is probably "and are those terms actually legally enforceable" which ratchets things up another level. Depending on their jurisdiction, how they did things, and what exactly Nvidia tried to go after it could be anything from illegal to gray to not. I'd lean towards the latter end of that spectrum, even if the EULA had some scare terms around it, but I'm also just some dude on the internet who hasn't even read the exact terms in the EULA or looked into if they'd be relevant to how ZLUDA was built/used.

Re: Zluda 6 release (run unmodified CUDA applications on non-Nvidia GPUs)

#13

The 32 bit PhysX support is particularly interesting as Nvidia had wanted to remove support for that particular combo on the 5000 series. They eventually flipped back to supporting it due to the pushback, but there was a short period where people were wondering if the same kind of 3rd party translation layer would be needed for Nvidia's own GPUs in this specific scenario.

it's emulated on the 5000 series right?

Re: Zluda 6 release (run unmodified CUDA applications on non-Nvidia GPUs)

#14

Isn't this tool directly violating NVIDIA license terms? I thought they disallow running CUDA on non-NVIDIA hardware. Or is it just because there's no money for NVIDIA in suing poor, Eastern European programmer?

Oracle v. Google found that copying API details sufficient to implement compatibility was not infringement.

Re: Zluda 6 release (run unmodified CUDA applications on non-Nvidia GPUs)

#15
post #4

I appreciate the focus on amusement in: > Some of the newly added features may come as a surprise to those of you who keep a close track of ZLUDA development. Most of them were previously explicitly outside ZLUDA's roadmap. There has been a change of plans. ZLUDA development is no longer commercially funded, so it's back to being my weekend project. This means that the priority is no longer what makes commercial sens…

it would be quite funny if they used the hip cuda backend to make new nvidia gpus run 32bit physx
Post reply on HN