Rackham#
Rackham is a high performance computer cluster at UPPMAX (Uppsala Multidisciplinary Center for Advanced Computational Science). UPPMAX is Uppsala University's resource of high-performance computers, large-scale storage and know-how of high-performance computing (HPC).
UPPMAX modules#
| Command | Comment |
|---|---|
| module avail | List available modules |
| module load modulename | Load the module modulename |
| module unload modulename | Load the module modulename |
| module list | List all modules loaded |
| module spider searchword | Search all module containing searchword |
Running jobs with the Slurm ressource manager#
| Command | Comment |
|---|---|
| interactive -A projectID | Start interactive job using 1 core for 1 hour |
| interactive -A projectID -t 8:00:0 -p core -n 2 | Start interactive job using 2 cores for 8 hours |
| sbatch -A projectID -t d-hh:mm:ss -n cores -p partition my_jobscript_file | Start batch job |
| sbatch -A projectID -t 7-00:00:00 -p node my_jobscript_file | Running for 7 days on all 20 cores one one node |
| sbatch -A projectID -t 48:00:00 -p core -n 4 my_jobscript_file | Running for 2 days on only 4 cores on a node |
| scancel jobId | Cancel a single job |
| scancel -i -u user | Interactively cancel all jobs for user |
Showing user, job and project info#
| Command | Comment | Full documentation |
|---|---|---|
| jobinfo | Show all running and waiting jobs in the queue | jobinfo |
| jobinfo -u user | Show jobs for specific user | jobinfo |
| jobstats -p -A projectID | generate plots for resource usage for recent jobs in a project | jobstats |
| uquota | Show disk usage of projects you are a member of | |
| projinfo | Show used core hours of projects you are a member of | |
| finishedjobinfo | Telling you about finished jobs on Rackham | |
| projmembers | List members of a project | |
| projsummary projectID | Summarizes some useful information about projects | |
| squeue -u user | Information about user's jobs |
/!\ Most of the commands use the cluster you are logged into as default cluster, if you wish to apply the command to another cluster you have to specify is using:
-M clusterName
/!\ Most of the commands have plenty of parameters and have an help. So don't forget to check it using:
command --help