donaldchung-hk/phonopy-castep-slurm-job-generator

Phonopy-CASTEP-Slurm-Job-Generator

A python script to generate and submit a batch of slurm job

This should be use as a more sophisicaed bash script as you can make changes to various part of the .py and .template file to suit your need

Requirements

• phonopy: 2.17.1 from conda-forge (This is the only tested system)
• CASTEP21 compiled version run usuing mpirun castep.mpi
• You need a output file from an optimisation run from CASTEP21 captured using {seed_name}-out.cell
• (Optional) You may want to run the simulation using SYMMETRY_GENERATE
• (Optional) euphonic: 1.2.0 from conda-forge

To run

1. Edit the parameter section in the .py file.
2. Edit the template files to suit your needs. Refer to Python3 - Template strings on how to set this up.
3. Edit the *_dict objects of the corresponding template if you make changes to the template that changes the variables.
4. You can either test run one of them or run it on a remote using source submit.sh .
5. Results are collected to ./result folder.

Analysis

This part is from personal project but can be adapted for other research as well.

1. Copy the phonopy_disp.yaml and the optimised output to the result folder
2. cd to the result folder and run phonopy --castep -f {system_name}*.castep --include-all to get the force (change the {system_name} as it is a variable in the script)
3. Run phonopy --castep -c {path to optimised structure} --dim="{your_kpoint_dimension}" --dos --mesh="{your_mash size}" --include-all to get the dos
4. Run euphonic-dispersion phonopy.yaml to plot and view the DOS.