====== 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** (infos wrt. running jobs) - **squeue** (show queues) - sinfo (show node infos) ==== STUBL Tools ==== - snodes (list node features; wrapper for sinfo) - sqstat (list job queues and cluster status information) \\ ===== Manpages ===== * ''man 1 sbatch'' * ''man 1 srun'' * ''man 1 salloc'' * ''man 1 scancel'' * ''man 1 squeue'' * ''man 1 sinfo'' \\ ===== Important switches for sbatch/srun/salloc ===== - ''-a '' job array with index range, e.g. ''-a 0,1,4-6'' or ''-a 0-8:2'' (environment variables available: **SLURM_ARRAY_TASK_ID**, **SLURM_ARRAY_JOB_ID**, **SLURM_ARRAY_TASK_COUNT** etc.) - '' -d '' define dependencies on other jobs (dependency_list: **after:job_id[:job_id]** or **afterok:job_id[:job_id]** and so on) - --''begin='' (absolute/relative) start time of job (see ''man 1 sbatch'') - --''deadline='' job only scheduled as long as it finishes before deadline - --''mem='' amount of **virtual memory per node** to allocate, e.g. 1G (1 Gibibyte) - ''-t '' set job's wall clock time - ''-e '' where to redirect error output - ''-o '' where to redirect stdout - ''-i '' read stdin for job script from file - ''-J '' - ''-n '' the overall amount of task (all nodes); does NOT accept range of task number :-( - ''-N '' overall number of nodes; accepts range of node number :-D - --tasks-per-node= number of tasks launched on each node - ''-w '' where e. g.: node001; node[001-005,009]; node001,node008 - --''mail-type='' **NONE**, **BEGIN**, **END**, **FAIL**, **REQUEUE**, **ALL**, **ARRAY_TASKS** etc., for further explanations see ''man 1 sbatch'' - ''-v'' run in verbose mode \\ ===== Useful combination of arguments for srun/sbatch/salloc ===== - -N with -n - -N with -n and --tasks-per-node= \\ date of revision: 10-30-2018 © kraus