Nvidia burnished its open source credentials this week after buying the company behind the veteran Slurm scheduler and announcing a slew of open source AI models.
The chip giant revealed yesterday that it had acquired SchedMD, the key developer behind Slurm, which Nvidia described in a statement as “an open source workload management system for high-performance computing (HPC) and AI.”
Slurm has been around since 2002, when backers included Lawrence Livermore National Laboratory and France’s Groupe Bull.
Nvidia said this week it will “continue to develop and distribute Slurm as open source, vendor-neutral software, making it widely available to and supported by the broader HPC and AI community across diverse hardware and software environments.”
It added that “Nvidia will accelerate SchedMD’s access to new systems – allowing users of Nvidia’s accelerated computing platform to optimize workloads across their entire compute infrastructure.” But Nvidia insisted it will also support “a diverse hardware and software ecosystem, so customers can run heterogeneous clusters with the latest Slurm innovations.”
Nvidia further played the open card, with the debut of its Nemotron 3 family of models, which it said delivered a “hybrid latent mixture-of-experts (MoE)” architecture in Nano, Super, and Ultra sizes.
The firm also said the releases support its broader “sovereign AI efforts.”
Nemotron 3 NANO is a 30 billion-parameter model, activating up to 3 billion parameters at a time for “targeted” tasks. Super is a “high-accuracy reasoning model” with approximately 100 billion parameters for multi-agent applications. Ultra comes in at 500 billion parameters for “complex AI applications.” Nvidia said the new architecture means the NANO model achieves four times the token throughput compared to its predecessor.
Nvidia has regularly been accused of locking customers into its ecosystem, as much through its CUDA software platform as its pricey silicon.
But it has increasingly sought to portray itself as an open champion. Earlier this year, it open sourced the Kubernetes native KAI scheduler it inherited through its purchase of Run:AI. At GTC in March, it open sourced its cuOpt optimization engine and its Groot robotics model.
There’s open and “open,” however, and observers suggest the SchedMD buy inevitably strengthens Nvidia’s grip on the overall stack, with expectations that its silicon will have a speed advantage when running the tool. ®
