NVIDIA acquires developer of job scheduler 'Slurm,' strengthening AI and HPC ecosystem

NVIDIA has announced that it has acquired

https://blogs.nvidia.com/blog/nvidia-acquires-schedmd/

SchedMD's main product is the job scheduler ' Slurm .' Slurm is responsible for appropriately allocating resources such as CPU and memory in environments where many users are using it simultaneously, and is used by more than 65% of the systems ranked in the ' TOP500 ' supercomputer performance ranking.

Slurm is developed as open source and has the following features: it can be used on any cluster, regardless of size; it excels at GPU resource management; it uses a plug-in-based architecture that makes it easy to adapt to organizational needs; and it can be used both on-premise and in the cloud.

NVIDIA says of Slurm, 'Slurm is used by AI developers to manage the resources needed for training and inference, and is also part of the critical infrastructure required for generative AI.'

NVIDIA has been working with SchedMD for over 10 years. Following the acquisition, the company plans to accelerate access to SchedMD's new systems, expand workload optimization and hardware and software support on NVIDIA's HPC platform, and enable users to run heterogeneous clusters. NVIDIA also stated that it will keep Slurm open source even after acquiring SchedMD.

'This acquisition is a testament to Slurm's importance in the world's most demanding HPC and AI fields,' said Danny Aubre, CEO of SchedMD. 'NVIDIA's deep expertise and investment in accelerated computing will strengthen Slurm's continued open source development and meet the demands of next-generation AI and supercomputing.'

The Slurm documentation is available at the following link:

Slurm Workload Manager - Documentation
https://slurm.schedmd.com/