Unterschiede

Hier werden die Unterschiede zwischen zwei Versionen angezeigt.

--- hpc:scheduling:slurm_commands [2024/06/13 18:16] – ↷ Seite von hpc:tutorials:scheduling:slurm_commands nach hpc:scheduling:slurm_commands verschoben
+++ hpc:scheduling:slurm_commands [2024/10/25 15:03] (aktuell) –
@@ Zeile 1: / Zeile 1: @@
 ==== Slurm Commands ====
-=== Job starten ===
+=== Start Jobs ===
+**srun**
+This is the simplest way to run a job on a cluster.
+Initiate parallel job steps within a job or start an interactive job (with --pty).
+**salloc**
+Request interactive jobs/allocations.
+When the job is started a shell (or other program specified on the command line) it is started on the submission host (Frontend).
+From this shell you should use srun to interactively start a parallel applications.
+The allocation is released when the user exits the shell.
+**sbatch**
+Submit a batch script.
+The script will be executed on the first node of the allocation.
+The working directory coincides with the working directory of the sbatch directory.
+Within the script one or multiple srun commands can be used to create job steps and execute parallel applications.
+**Examples**
 <code>
-sbatch --job-name=$jobname -N <num_nodes> --ntasks-per-node=<ppn> Jobscript
+# General:
-Eine Startzeit kann mit dem Schalter --begin vorgegeben werden
+sbatch --job-name=<name of job shown in squeue> -N <num nodes> --ntasks-per-node=<spawned processes per node> /path/to/sbatch.script.sh
-Beispielsweise:
+# A start date/time can be set via the --begin parameter:
 --begin=16:00
 --begin=now+1hour
 --begin=now+60 (seconds by default)
 --begin=2010-01-20T12:34:00
-Weitere Informationen bietet auch man sbatch
-Sämtliche dort verwendeten Parameter können auch im Jobscript selber mit #SBATCH angebenen werden.
 </code>
-=== Status eigener Jobs abfragen ===
+A sbatch script of the command above would look like
 <code>
-squeue
+#!/bin/bash
+#SBATCH --job-name=<name of job shown in squeue> -N <num nodes>
+#SBATCH -N <num nodes>
+#SBATCH --ntasks-per-node=<spawned processes per node>
+#SBATCH --begin=2010-01-20T12:34:00
+/path/to/sbatch.script.sh
 </code>
-=== Knotenstatus anzeigen ===
+For more information see ''man sbatch''.
+All parameters used there can also be specified in the job script itself using #SBATCH.
+Also, more examples can be found [[hpc:tutorials:sbatch_examples|here]].
+=== Check the status of your job submissions ===
+<code>
+squeue --me
+</code>
+=== Check the status of nodes ===
 <code>
 sinfo
 </code>
-=== Jobs löschen ===
+=== Canceling jobs ===
+On allocation, you will be notified of the job ID.
+Also, within your scripts and shells (if allocated via salloc) you can get the ID via the ''$SLURM_JOBID'' environment variable.
+You can use this ID to cancel your submission:
 <code>
 scancel <jobid>
 </code>
+=== Slurm Cheat Sheet ===
+A summary of the most common commands can be found [[https://slurm.schedmd.com/pdfs/summary.pdf|here]].