| Slurm | |
|---|---|
 | |
| Developer | SchedMD |
| Stable release | |
| Repository | |
| Written in | C |
| Operating system | Linux |
| Type | Job Scheduler for Clusters and Supercomputers |
| License | GNU General Public License |
| Website | slurm |
TheSlurm Workload Manager, formerly known asSimple Linux Utility for Resource Management (SLURM), or simplySlurm, is afree and open-sourcejob scheduler forLinux andUnix-likekernels, used by many of the world'ssupercomputers andcomputer clusters.
It provides three key functions:
Slurm is the workload manager on about 60% of theTOP500 supercomputers.[2]
Slurm uses abest-fit algorithm based onHilbert curve scheduling orfat tree network topology in order to optimize locality of task assignments on parallel computers.[3]
Slurm began development as a collaborative effort primarily byLawrence Livermore National Laboratory,SchedMD,[4] Linux NetworX,Hewlett-Packard, andGroupe Bull as a Free Software resource manager. The first release happened in 2002.[5] It was inspired by the closed sourceQuadrics RMS and shares a similar syntax. The name is a reference to thesoda inFuturama.[6] Over 100 people around the world have contributed to the project. It has since evolved into a sophisticated batch scheduler capable of satisfying the requirements of many large computer centers.
As of November 2021[update],TOP500 list of most powerful computers in the world indicates that Slurm is the workload manager on more than half of the top ten systems.
Slurm's design is very modular with about 100 optional plugins. In its simplest configuration, it can be installed and configured in a couple of minutes. More sophisticated configurations provide database integration for accounting, management of resource limits and workload prioritization.
Slurm features include:[7]
Recent Slurm releases run only onLinux. Older versions had been ported to a few otherPOSIX-basedoperating systems, includingBSDs (FreeBSD,NetBSD andOpenBSD),[8] but this is no longerfeasible as Slurm now requirescgroups for core operations. Clusters running operating systems other than Linux will need to usea different batch system, such as LPJS.[9] Slurm also supports several unique computer architectures, including:
Slurm is available under theGNU General Public License v2.
In 2010, the developers of Slurm founded SchedMD, which maintains the canonical source, provides development, level 3 commercial support and training services. Commercial support is also available fromBull,Cray, and Science + Computing (subsidiary ofAtos).
Theslurm system has three main parts:
slurmctld, a central controldaemon running on a single control node (optionally withfailover backups);slurmd daemons;The clients can issue commands to the control daemon, which would accept and divide the workload to the computing daemons.
For clients, the main commands aresrun (queue up an interactive job),sbatch (queue up a job),squeue (print the job queue) andscancel (remove a job from the queue).
Jobs can be run inbatch mode orinteractive mode. For interactive mode, a compute node would start a shell, connects the client into it, and run the job. From there the user may observe and interact with the job while it is running. Usually, interactive jobs are used for initial debugging, and after debugging, the same job would be submitted bysbatch. For a batch mode job, itsstdout andstderr outputs are typically directed to text files for later inspection.
