In the short term this changes little, the should be build-able similar to how the other proprietary NV kernel modules are now though. In time they'll no doubt go upstream, either in their own right or be integrated into the nouveau driver. This will take time and I suspect the outcome will be very similar to the way the AMD GPU drivers work where they have an open driver and a firmware. Overall it's good news and positive for all of the ecosystem. It seems also it's primarily focused, at least initially, on their data centre GPUs so I suspect they'll be more stable/useful for AI/ML/CUDA workloads than the average desktop user. > As far as I know this would still not be acceptable in Fedora due to the "No > external kernel modules" rule, right? > > https://docs.fedoraproject.org/en-US/packaging-guidelines/what-can-be-pac... > > -- > Miro Hrončok > -- > Phone: +420777974800 > IRC: mhroncok > _______________________________________________ > devel mailing list -- devel(a)lists.fedoraproject.org > To unsubscribe send an email to devel-leave(a)lists.fedoraproject.org > Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ > List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines > List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org > Do not reply to spam on the list, report it: https://pagure.io/fedora-infrastructure

I don't think Nvidia has any intention to add this to mainline kernel as it would block use of the propriety driver. I zero interest in packaging it for rpmfusion as it's so limited, see http://us.download.nvidia.com/XFree86/Linux-x86_64/515.43.04/README/kerne...

On 12/05/2022 11:40, Aleksandra Fedorova wrote: This will only be possible if NVIDIA releases GPU specs. -- Sincerely, Vitaly Zaitsev (vitaly(a)easycoding.org)

That would block the use of 390xx, 470xx and mai, see http://us.download.nvidia.com/XFree86/Linux-x86_64/515.43.04/README/kerne...