Pipeline with scripts for running pulsar search software (PRESTO and custom Python routines) on HPC nodes.
HPC_pulsar
is written in Python 3, tested on Python 3.7.4 and has the following package dependencies.
The above packages have several underlying dependencies that are listed in their respective repositories.
Note: To avoid the hassle of installing PRESTO
and its several dependencies, you can pull a Singularity container encasing a full PRESTO
installation here via the following command line call.
singularity pull shub://federatedcloud/singularity-PRESTO
Alternately, for a Singularity container with both PRESTO
and [psrdynspec](https://github.com/akshaysuresh1/psrdynspec)
installed, run the below command line code.
singularity pull shub://federatedcloud/ pulsar-pipeline-container
Batch scripts found the batch_HPC
directory provide example command line calls run within a singularity container.
batch_HPC
: Batch scripts for non-interactive execution via Slurm.cmd_files
: Command line calls invoked by batch scripts for non-interactive program run.config
: Configuration script of inputs to different modules.executables
: Executable scripts that perform the computation. NOTE: Do not tamper with these scripts.
- Clone this repository to your local machine.
- Edit relevant files in the
batch_HPC
,cmd_files
andconfig
folders. - Pass relevant batch script to Slurm. You can now wait and relax while your code execution progresses non-interactively.
Say that you want to dedisperse some data using MPI-enabled prepsubband
of PRESTO
.
- Edit inputs in the configuration script
config/dedisp.cfg
. - Verify that the paths to the relavant executable file and config script are correctly provided in the Shell script
cmd_files/dedispersion.cmd
. - Supply appropriate parameters to Slurm in
batch_HPC/dedispersion.sh
. - Type
sbatch batch_HPC/dedispersion.sh
and hit Enter.
Please submit an issue to voice any problems or requests.
Improvements to the code are always welcome.