ICC Home Privacy and Legal Notice LC User Documents Banner

UCRL-WEB-201386

SLURM Reference Manual


SLURM Operation

Links to subtopics in this section:
  • SLURM Utilities

  • SRUN (Submit Jobs)

  • SQUEUE (List Jobs)

  • SINFO (List Nodes)

  • SMAP (Show Job Geometry)

  • SCONTROL (Manage Configurations)

  • SCANCEL (Signal or Kill Jobs)


  • Navigation Links: [ Document List ] [ HPC Home ] [ Previous ]

    No content in this section. Showing first subtopic...

    SLURM Utilities

    SLURM's six command-line utilities provide its direct interface for users (while LCRM and Moab utilities, as explained in EZJOBCONTROL, provide an indirect interface). These utilities are:

    SRUN
    submits jobs to run under SLURM management. SRUN can
    (A) submit a batch job and then terminate, or
    (B) submit an interactive job and then persist to shepherd the job as it runs, or
    (C) allocate resources to a shell and then spawn that shell for use in running subordinate jobs.
    SLURM associates every set of parallel tasks ("job steps") with the SRUN instance that initiated that set, and SRUN gives you elaborate control over node choice and I/O redirection for your parallel job. (Job steps are not supported on BlueGene/L.)
    SQUEUE
    displays (by default) the queue of running and waiting jobs (or "job steps"), including the JobId (used for SCANCEL), and the nodes assigned to each running job. But you can customize SQUEUE reports to cover any of 24 different job properties, sorted by the properties most important to you.
    SINFO
    displays a summary of status information on SLURM-managed partitions and nodes (not jobs). Customizable SINFO reports can cover the node count, state, and name list for a whole partition, or the CPUs, memory, disk space, or current status for individual specified nodes.
    SCANCEL
    cancels a running or waiting job, or sends a specified signal to all processes on all nodes associated with a job (only job owners or their administrators can cancel their jobs).
    SCONTROL
    (privileged users only) manages available nodes (for example, by "draining" jobs from a node or partition to prepare it for servicing) and assigns properties to node partitions.
    SMAP
    (on BlueGene/L only) displays a character-based chart or plot showing how nodes are allocated geometrically among current jobs and BG/L partitions (a job-planning tool).


    Navigation Links: [ Document List ] [ HPC Home ] [ Next ]