Nvidia is publishing its Linux GPU kernel modules based on the open source dual GPL / MIT license now available on GitHub.
This is a good day for Linux Dystros users because the hardware giant allows this action to be strictly integrated with the OS and allows developers to debug, integrate and re-collaborate with their communities.
The move follows a period of colorful history between Linux and Nvidia, with Linus Torvalds, chief architect of the Linux kernel, once calling Nvidia “the worst company we’ve ever dealt with” in a 2012 speech.
What does this mean for consumers?
This action allows developers to identify code paths and see how kernel event scheduling interacts with their workload, enabling faster elementary debugging, Nvidia claims.
Additionally, Nvidia claims that enterprise software developers can now integrate the driver into a customized Linux kernel configured for their project.
It helps improve the quality and security of Nvidia GPU drivers through input and reviews from the Linux end user community, according to the company.
Support, however, is not set to be shared equally across all of Nvidia’s hardware.
While NVIDIA Turing and NVIDIA Ampere architecture families are ready to generate code for its data center GPUs, while support for GeForce and Workstation GPUs is only of alpha quality, this latest announcement is less relevant to laptop and consumer users.
Nvidia has promised to introduce more GeForce and Workstation support in future releases, and NVIDIA OpenKernel modules will eventually replace the closed-source driver.
The open-source kernel-mode driver works with the same firmware and the same user-mode stock as Nvidia with some other offers, including CUDA, OpenGL and Vulkan.
However, all components in the driver stack must match the versions in the release.
If you are interested in trying out new kernel modules, go to Driver Read on Documentation for instructions on installing the correct versions and for additional troubleshooting steps.