Nouveau Finally Tames NVIDIA’s GA100 in Linux 7.2 Kernel

Linux 7.2 brings long-awaited Nouveau support for NVIDIA's compute-only GA100 GPU via GSP firmware. The kernel-level enablement offers an open alternative to proprietary drivers, though user-space acceleration remains limited without a 3D engine. This change expands options for data center operators.
Nouveau Finally Tames NVIDIA’s GA100 in Linux 7.2 Kernel
Written by Sara Donnelly

Years of waiting end this week for data center operators running NVIDIA’s compute-focused GA100 chips on Linux systems. The open-source Nouveau driver gains official support for the GPU in the upcoming Linux 7.2 kernel. This change arrives courtesy of the final drm-misc-next pull request ahead of the merge window. And it marks a quiet but meaningful step forward for those who prefer to avoid proprietary binaries.

The GA100, part of NVIDIA’s Ampere architecture, stands apart as a compute-only accelerator. It lacks the 3D graphics engine found in consumer counterparts. That distinction kept it from full Nouveau integration until now. Phoronix reported in February that NVIDIA engineers posted patches to enable the card through the GPU System Processor, or GSP. Those patches treated the GA100 much like a Turing chip for bring-up purposes. They relied on Turing GSP-RM firmware.

But why the delay? Earlier Ampere cards had gained GSP support in Nouveau. The GA100 required specialized handling. Its compute-only design meant different code paths. With the patches merged, the kernel can now initialize the GPU. Provided, of course, that the necessary firmware files sit in place. Short, simple requirement. Yet one that opens doors for servers where every accelerator counts.

Don’t expect miracles on day one. User-space components still lag. Both Clover and Rusticl or NVK in Mesa need a 3D engine to deliver meaningful acceleration. Without it, the GA100 sees limited action beyond basic kernel recognition. “Although Nouveau now supports GA100, there is currently no real support for this GPU with either Clover or Rusticl/NVK, as both require a 3D engine,” noted the February coverage. This contrasts with Hopper GPUs like the GH100, which include a graphics engine and enjoy better upstream readiness.

The development reflects broader shifts in how NVIDIA interacts with the Linux community. Once reluctant, the company now contributes directly to Nouveau. These patches represent one more piece in that pattern. They reduce dependence on out-of-tree modules for certain enterprise hardware. Server administrators gain another option when deploying fleets of A100-based systems. Performance remains unknown in open source for now. But basic functionality beats none at all.

News of the merge broke late on May 28. Michael Larabel detailed the pull request at Phoronix. It wasn’t the only change. The same batch added a DSC quirk for the ASUS DC301 USB-C dock and assorted small fixes across DRM drivers. Routine housekeeping alongside the headline GPU work. Still, the GA100 addition stands out for its implications in high-performance computing circles.

Context matters here. The GA100 powered many early A100 accelerators that dominated AI training and scientific workloads several years ago. Those cards remain in service. Some organizations run them in clusters where stability and openness carry weight. Others simply want to avoid the closed driver for philosophical or operational reasons. Nouveau’s progress gives them breathing room. Not full parity. But a foundation.

Related efforts continue elsewhere. Mesa 25.2 expanded NVK Vulkan support to Blackwell and older Kepler parts, according to a Collabora announcement. Such updates show the open NVIDIA stack gaining breadth even as gaps persist for specific chips like the GA100. Hopper received theoretical enablement there too, though real hardware testing stays limited.

So what comes next? Kernel support is necessary but not sufficient. Developers must extend compute interfaces. OpenCL, CUDA alternatives, or direct kernel APIs could unlock the card’s potential without graphics. That work sits ahead. In the meantime, Linux 7.2 users with GA100 cards can test the new code. Load the module. Check dmesg output. Watch for successful initialization. The experience will vary by firmware and system configuration.

This milestone arrives amid ongoing conversations about NVIDIA’s role in open source. The company ships its own kernel modules. Yet contributions like these GSP patches signal cooperation. They help fill holes that proprietary drivers sometimes ignore in older or niche configurations. Enterprise users notice. They value long-term support without vendor lock-in risks.

Expect more refinements before Linux 7.2 ships. The merge window has only begun. Additional fixes could refine GA100 behavior. Power management, memory handling, or error reporting might see attention. The community will scrutinize results from early testers. Feedback will shape what lands in the final release.

For now the news feels understated. No fanfare. Just code moving from mailing list to mainline. But for operators managing racks of aging Ampere accelerators, it matters. They gain a fully upstream path. They avoid proprietary blobs where possible. And they participate in an ecosystem that values transparency. Small victories accumulate. This one counts.

Subscribe for Updates

DevNews Newsletter

The DevNews Email Newsletter is essential for software developers, web developers, programmers, and tech decision-makers. Perfect for professionals driving innovation and building the future of tech.

By signing up for our newsletter you agree to receive content related to ientry.com / webpronews.com and our affiliate partners. For additional information refer to our terms of service.

Notice an error?

Help us improve our content by reporting any issues you find.

Get the WebProNews newsletter delivered to your inbox

Get the free daily newsletter read by decision makers

Subscribe
Advertise with Us

Ready to get started?

Get our media kit

Advertise with Us