> 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 sense, but what I find the most entertaining. That's why the sudden addition of textures, PhysX and better Windows support.
Outside of the time commitment that full-time support would offer, I generally think focusing on amusement is a good life strategy.
Agreed. I think the real value of Zluda to the community is going to lie on more niche applications. The ones that are more commercially interesting are likely to have other solutions developed for them.
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.
Iirc he was kind-of funded by AMD for a while to break the CUDA LLM moat (the mention of change of direction at the bottom of the article), no idea if it was legal issues with him being paid by them or why it changed, but..
1: Much initial research and early OSS was CUDA focused due to libaries and momentum, but with more Mac users,etc kernels/frameworks are more portable so with stuff like Unsloth so the CUDA moat is falling there (If it's Vulkan or more propietary backends doesn't matter, Vulkan or OpenCL are probably fairly equal in usefulness since Vulkan support would probably require custom extensions for full performance).
2: As alluded above, ZLuda seems to have focused a bit on ROCm API's for AMD (I think I've seen mentions of other library supports but it's been up and down iirc since he was employed by AMD). The main benefit compared to Vulkan iirc is that ROCm allows for some CUDA-like feature that makes kernel emulation so much less troublesome (or even feasible at all?).
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.
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 sense, but what I find the most entertaining. That's why the sudden addition of textures, PhysX and better Windows support.
Outside of the time commitment that full-time support would offer, I generally think focusing on amusement is a good life strategy.
Agreed. I think the real value of Zluda to the community is going to lie on more niche applications. The ones that are more commercially interesting are likely to have other solutions developed for them.
it would be quite funny if they used the hip cuda backend to make new nvidia gpus run 32bit physx
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?
Nice pun
Złuda means "mirage"/"illusion"/"figment of imagination"
Even funnier that CUDA in polish means “miracles”
Has anyone tried this for LLMs and how does it compare to Vulkan?
Iirc he was kind-of funded by AMD for a while to break the CUDA LLM moat (the mention of change of direction at the bottom of the article), no idea if it was legal issues with him being paid by them or why it changed, but..
1: Much initial research and early OSS was CUDA focused due to libaries and momentum, but with more Mac users,etc kernels/frameworks are more portable so with stuff like Unsloth so the CUDA moat is falling there (If it's Vulkan or more propietary backends doesn't matter, Vulkan or OpenCL are probably fairly equal in usefulness since Vulkan support would probably require custom extensions for full performance).
2: As alluded above, ZLuda seems to have focused a bit on ROCm API's for AMD (I think I've seen mentions of other library supports but it's been up and down iirc since he was employed by AMD). The main benefit compared to Vulkan iirc is that ROCm allows for some CUDA-like feature that makes kernel emulation so much less troublesome (or even feasible at all?).
Do other models - VLMs, Diffusion, etc., also work fine with Vulcan?
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.
Oracle v. Google found that copying API details sufficient to implement compatibility was not infringement.
Useful context https://www.tomshardware.com/pc-components/gpu-drivers/cuda-...
[flagged]