Skip to content

Running jobs on HPC cluster

Scripts for compilation and running of TCLB solver on HPC clusters are avaiable on:


The scripts are made to be general, but HPC clusters are very diverse. Clusters with SLURM and PBS queue system are supported. The goal is to give a feeling of local execution, like:

wrapper/run solver case.xml

The batch files for SLURM/PBS are prepared and scheduled by the scripts. Currently, the setup of modules is prepared for following clusters:

  • gadi (National Computational Infrastructure, Australia)
  • hyperion (Warsaw University of Technology, Poland)
  • magnus (Pawsey Supercomputing Centre, Australia)
  • prometheus (Cyfronet, Poland)
  • rysy (ICM, University of Warsaw, Poland)
  • goliath


Getting the source

git clone
git clone p


All the needed configuration and installation of R packages is done by:


Hit enter to agree on defaults.


As computation on the main node is discouraged on many clusters, you can run make in parallel on a computational node.

p/make d2q9

Running cases

To run cases you can call the run script:

p/run [model] [case.xml] [number of cores/gpus] 

You can use the script from a different directory then TCLB main dir, so for instance:

cd /scratch/blarbla/blarbla/some_important_research/
~/TCLB/p/run d2q9 mycase.xml 

Changeing options

You can change the default number of mpi_processes/gpus (1), etc by supplying additional options for sbatch like:

~/TCLB/p/run d2q9 mycase.xml 4 

Other flags are:

--print # dry run - to see the batch file
--wait # to see the output on-the-fly in terminal


The scripts support getting PushBullet notifications based on:

Cluster specific settings

In the cluster directory there are settings for specific clusters (like prometheus and magnus). The first line of each file is a bash comment with the pattern for hostname to match.


If you want to add support for another cluster, please add another file in the cluster directory.