IT-Bereich des Instituts für Chemie hpc:hpc_tubit:slurm_usage

User/Group Accounting

Anonymous (anonymous@undisclosed.example.com) — 2018-09-10T20:08:50+00:00

User/Group Accounting * sacct -g quantenchemie_group (lists all recent jobs) * sacct -g quantenchemie_group -s R (lists all running jobs) * sacct -u $USER -s R * seems to be no option for XML output :-( (e.g. use awk to sum up) date of revision: 09-10-2018 © kraus

Batch Jobs

Anonymous (anonymous@undisclosed.example.com) — 2020-02-26T10:56:46+00:00

Batch Jobs * Equeueing of scripts launched via: sbatch job_script * Please DO NOT employ the parameter --ntasks-per-core= unless you really know what you do. Erroneous usage of this flag
(especially in combination with the parameter

Useful commands

Anonymous (anonymous@undisclosed.example.com) — 2019-06-24T13:18:31+00:00

Useful commands * Showing information about status and resources of each host within the cluster: sinfo -Nel * Showing your own job queue: sqeueue -u$USER * Showing the queue's status on a specific partition: squeue -p[partition] * Sending a SLURM script to the queue:

Common terms in SLURM language

Anonymous (anonymous@undisclosed.example.com) — 2020-02-26T14:07:04+00:00

Common terms in SLURM language * cpu: hardware processor, generally termed core (not logical CPU entity, no hyperthreading) * node: computing host * queue: scheduling of jobs on a partition/group of nodes * partition: a group of nodes with similar computing capabilities

External APIs

Anonymous (anonymous@undisclosed.example.com) — 2018-09-10T20:09:08+00:00

External APIs Python * A general Python interface exists: and (seen on 09-10-2018) date of revision: 09-10-2018 © kraus

Commands

Anonymous (anonymous@undisclosed.example.com) — 2018-10-30T18:15:58+00:00

Commands Common Commands * slurm (introduction) * launching jobs: salloc (resource reservation for jobs) && srun (jobs and job steps running tasks) * batch jobs: sbatch (batch jobs to queue) * scancel (signal/stop running jobs) * sstat

Interactive srun jobs

Anonymous (anonymous@undisclosed.example.com) — 2022-10-06T17:38:20+00:00

Interactive srun jobs Standard nodes: srun -pstandard -t20:00 --mem=2G -n2 -N1 --hint=nomultithread --pty /bin/bash -il SMP nodes: srun -psmp -t20:00 --mem=2G -n2 -N1 --hint=nomultithread

Memory Limits

Anonymous (anonymous@undisclosed.example.com) — 2019-06-24T13:07:51+00:00

Memory Limits Type of Memory Limits * Resident memory (RSS): amount real memory in use by process * Virtual memory (VMem): amount of virtual memory pages allocated by process * In the current cluster setup, only the limit for the real memory has to be taken into consideration for jobs (by parameter

Example batch job scripts for program packages

Anonymous (anonymous@undisclosed.example.com) — 2019-05-02T15:56:14+00:00

Example batch job scripts for program packages packages_jobscript index

Links

Anonymous (anonymous@undisclosed.example.com) — 2018-10-20T21:33:44+00:00

Links * (external, seen on 10-20-2018) * * * * (external, seen on 08-12-2018) * * …

Using SLURM batch job facility

Anonymous (anonymous@undisclosed.example.com) — 2019-05-02T15:52:26+00:00

Using SLURM batch job facility slurm_usage index