CUDA, Meet SUSE. Your GPU Workflow Just Got Simpler.

Building GPU-accelerated applications can involve navigating the complexities of software dependencies, versions, and installations from various sources.  It’s a process that can be challenging and get in the way of what you actually want to do: build amazing things.

Today, we’re thrilled to announce that this is evolving.  

Following a close collaboration with NVIDIA, SUSE can now distribute the NVIDIA CUDA Toolkit directly within our products. You might have already seen the news from NVIDIA about this; we’re excited to share what this means for you, our developer community. Our goal is simple: to make deploying CUDA on SUSE platforms radically easier, helping you accelerate your work in AI, high-performance computing (HPC), and beyond.

CUDA, Natively on SUSE

We’ve teamed with NVIDIA to bring the CUDA software stack directly into SUSE products.

This means you can now get the essential CUDA components right alongside your other SUSE packages, which will streamline your entire setup and dependency management. This is a game-changer, especially for complex AI frameworks like PyTorch and essential libraries like OpenCV.

Here’s what you get:

• Simplified Installation: All the necessary software is available from a single location. CUDA components are now integrated, making your installations straightforward and much less prone to error.
• Predictable and Consistent: We are committed to using NVIDIA’s exact CUDA Toolkit naming conventions, so there’s no guesswork or confusion when you’re following documentation or tutorials.
• Stay Current, effortlessly: SUSE will provide timely updates to CUDA packages right after NVIDIA’s official releases. This helps you stay compatible with the latest features and reduces your own quality assurance workload.
• CUDA Stays Free: Let’s be clear: while it’s part of our SUSE offerings, the NVIDIA CUDA Toolkit itself remains completely free to our users. You won’t need to buy another SUSE product just to get access to CUDA.
• We’ve Got Your Back: You’ll continue to get the reliable support you expect from SUSE for our integrated products. For deep, CUDA-specific questions, NVIDIA’s extensive developer forums are still your best resource.

Where Can I Get It?

Our first product to include this direct CUDA distribution is SUSE AI, which has been expressly approved by NVIDIA for this purpose.

But this is just the beginning. We know there is huge interest in having this capability everywhere, especially in the cloud and enterprise environments. We are already working to bring integrated CUDA distribution to other key products like SUSE Linux Enterprise Server (SLES) VM images and our container images (like SLE BCI). We’ll keep you posted as more products are approved by NVIDIA.

What’s Next?

This partnership with NVIDIA is a major step forward in our mission to make life easier for developers on SUSE platforms. By working closely with industry leaders like NVIDIA, we are making sure that the tools you rely on are accessible, consistent, and simple to use, no matter how you build your applications.

Stay tuned. We’re just getting started on empowering your next wave of GPU-accelerated innovation.