Slurm environmental variables

When launching a job, slurm retains information about the servers allocated, cores assigned, working directory from which the job was launched, among other things. Slurm passes this information to the job via environmental variables. The table below shows some variables commonly used within the shell scripts of batch jobs.

Commonly used Slurm environmental variables

Slurm Variable Name	Description	Example values
$SLURM_CPUS_ON_NODE	Number of total CPUs allocated on node	8,3
$SLURM_CPUS_PER_TASK	Number of cores per task. I.e., the value given to the -cpus-per-task or -c sbatch options. Not set unless one of those options given.	8,3
$SLURM_JOB_ID	Job ID	5741192
$SLURM_JOB_NAME	Job Name	myjob
$SLURM_JOB_NODELIST	Nodes assigned to job	cpusm[01-03]
$SLURM_JOB_NUM_NODES	Number of nodes allocated to job	2
$SLURM_LOCALID	Index to core running on within node	4
$SLURM_NODEID	Index to node running on relative to nodes assigned to job	0
$SLURM_NNODES	Deprecated. Same as SLURM_JOB_NUM_NODES	4
$SLURM_NTASKS	Total number of tasks. Number of total cores is given by SLURM_NTASKS times SLURM_CPUS_PER_TASK	11
$SLURM_PROCID	Index of task relative to job	0
$SLURM_SUBMIT_DIR	Directory from which the sbatch command was invoked	/work/user/wdir
$SLURM_SUBMIT_HOST	Hostname of the computer from which the command sbatch was invoked.	login01
$SLURM_TASKS_PER_NODE	This gives a comma-delimited list of integers representing the task per the node, using the same ordering as in SLURM_JOB_NODELIST. If consecutive nodes have the same task count, the integer will be followed by ‘(xN)’, so the example value is for 2 tasks on the first three nodes and 1 task on the fourth node.	2(x3),1

For more comprehensive details, please read Section “OUTPUT ENVIRONMENT VARIABLES” of Slurm’s sbatch manual.