Quick start guide

"Nothing is better than the wordless teaching and the advantage of non-action"


The cluster consist of the front-end and a number of computational nodes. The user submits a job from the front-end and the queueing system takes care to reserve requested resources and run the job when the time is optimal. An advanced job scheduling can assure fair access for everybody and optimize the overall system efficiency.



To use the cluster one has to 'ssh' to it's frontend node - psk. Each user of the CAMK network can use it - there is no separate account.
The frontend can be used to:

  • submit and monitor jobs
  • develop code (compiling; short debugging runs; profiling)


The frontend has 12 cpus and 32GB of memory. Please read the messages displayed after login - they contain current, important announcements. Do not run long or memory consuming codes for production on the frontend. There are certain limits set and the system will kill proceses which violate them.



There is one filesystem:

  • /work/psk - 140 TB volume, fast cluster filesystem - Lustre, default user quota 500GB, backup (mirror) done everyday around 5am;

It is recommended to use is for all cluster activities since all nodes and frontend have very fast link to it. The /work/psk filesystem is also accessible from all workstations at CAMK. Please note that the performance has priority over data safety here: use it to store e.g. simulation results but not for codes, papers and other data which can't be recovered.

Please send requests to create user catalog to psk@camk.edu.pl . Default quota is 500GB but can be increased if you need more.



Aside from the standard linux packages there is other software available:

  • Commercial PGI compilers (pgcc, pgc++, pgf77, pgf90, pgf95)
  • Software collections (devtoolset including newest gcc, python 3.3); to get current list: scl -l
  • MPI environments - have to be loaded with module command; to list options type: module available
  • Other, locally compiled software/libraries can be found in /opt directory

More details are in the PSK User's Guide.



The job is submited with command qsub script_filename. An example of the job script is shown below. Each line begining with #PBS is interpreted as an option to qsub command.

## specify job name
#PBS -N job_name
## destination queue
#PBS -q short
## request 1h of walltime
#PBS -l walltime=1:0:0
## request 200 MB of memory
#PBS -l mem=200mb
## email me if the job aborts or when completes
#PBS -m ae
## redirect standard output and standard error streams
#PBS -o out.txt
#PBS -e err.txt

## commands to execute
#go to the directory from which the job was submited (job starts in your /home directory)
#run code


To list available queues and their limits type qstat -q.
To list all jobs an their status type qstat.
To delete a job type qdel job_id (id is reported by qstat).
Please read 'man {qsub, qdel, qstat, pbs_resources}' for details.


More details on using the psk cluster can be foud in the PSK User's Guide. Please read it and check man pages before contacting support at psk©camk.edu.pl .