SLURM

SLURM (Simple Linux Utility for Resource Management) is the highly scalable, open-source workload manager that allocates compute nodes and schedules parallel jobs (e.g., MPI) across Linux clusters and supercomputers.

This is SLURM: the critical workload manager for high-performance computing (HPC) environments. It handles three core functions: allocating exclusive or non-exclusive access to compute nodes, providing a robust framework to start and monitor work (typically parallel jobs), and arbitrating resource contention by managing a queue of pending jobs. SLURM’s design is modular, fault-tolerant, and exceptionally scalable, proven to manage clusters with tens of millions of processors. It currently serves as the primary workload manager for approximately 60% of the world's TOP500 supercomputers: a clear metric of its reliability and performance in mission-critical operations.

https://slurm.schedmd.com