For more years than I care to remember, Linux users have hated NVIDIA, the world leader in Graphics Processing Units (GPUs). Why? Because years after every other company open sourced their drivers, NVIDIA refused to do so. It also refused to provide Linux kernel developers with the data they needed to build open source drivers for them. That left Linux users stuck with proprietary, second-rate drivers. Now, NVIDIA finally opens its GPU driver code.
Hate is not too strong a word. In a famous interview, Linus Torvalds said in frustration, “I’m also happy to publicly point out that Nvidia is one of the worst trouble spots we’ve ever had with hardware manufacturers, and it’s really sad because then Nvidia tries to sell chips – a lot of chips – into the Android Market. Nvidia is the worst company we’ve ever had to deal with. [Lifts middle finger] Therefore Nvidia, damn it.“
Also: Linux kernel 6.10 is official: Here’s what’s new and improved
And, let’s not forget, in 2022, the Lapsus$ ransomware gang threatened to leak Nvidia GPU files if the company refused” COMPLETELY OPEN SOURCE (and distributed under the foss license) their GPU driver for Windows, macOS, and Linux, now and forever.” This is no ordinary ransom demand.
It didn’t work, but NVIDIA is transitioning entirely to open-source GPU core modules as the company nears a $3 trillion market cap, marking a major shift from the company’s longstanding — and frankly, misguided — policy.
This transition began in May 2022 with the release of the R515 driver, which includes a set of Linux GPU kernel modules as open source under a dual GPL and MIT license. This release initially targets datacenter compute GPUs, with GeForce and Workstation GPUs in alpha status.
Over the past two years, NVIDIA has worked hard to improve these open source modules, achieving application performance equal to or better than their closed source counterparts. The company has also added significant new capabilities, including:
- Support for heterogeneous memory management (HMM)
- Secret calculation
- Support for NVIDIA’s Grace platform coherent memory architecture
These features won’t help much in achieving the best possible gaming performance, but they can help cryptocurrency miners or artificial intelligence (AI) developer.
Also: How Open Source Attracts Some of the World’s Top Innovators
With the upcoming release of the R560 driver, NVIDIA will complete its transition to open source GPU kernel modules. The move is expected to benefit the Linux community by allowing for better integration with the operating system and allowing for stronger community development and support.
It is important to note, however, that this transition does not apply to all NVIDIA GPUs. Open source modules are required for advanced platforms such as NVIDIA Grace Hopper or NVIDIA Blackwell. NVIDIA recommends switching to open source modules for GPUs from the Turing, Ampere, Ada Lovelace, or Hopper architectures. Older GPUs from the Maxwell, Pascal, or Volta architectures will continue to use proprietary drivers because they are not compatible with open source modules.
Need help figuring out what you have under the hood of your computer? NVIDIA provides a shell script, nvidia-driver-assistant, to help you. Once done, you can install CUDA Toolkit and the appropriate GPU driver using your Linux distribution’s package manager.
Also: Latest AI Training Benchmarks Show Nvidia Has No Competition
This move by NVIDIA is a step towards creating a more open ecosystem for Linux users and developers. It also puts NVIDIA in a better position to compete with AMD, which has long been known for its open-source-friendly approach to the Linux space.
While this is a significant development, it should be noted that NVIDIA’s transition to open source is currently limited to kernel modules. The user-space components of the driver are still proprietary, which means that full open source integration, especially for gaming applications, is still a long way off.
As the tech community digests the news, many hope it could be the start of a broader shift to open source operations at NVIDIA, potentially improving compatibility and performance for Linux users.
