How to use high-throughput migration calculation machine

Citation

If you are using the codes in this repo, please consider citing the following work:

Rong, Z., et al., An efficient algorithm for finding the minimum energy path for cation migration in ionic materials. The Journal of Chemical Physics, 2016. 145(7): p. 074112.

---

Setup external computational environment

1. Setup a personal MongoDB database. Update access confidentiality in settings/db_config.yaml
2. Install Materials Project MPWorks in the computation clusters to be deployed for the HT calculations.
3. Install the following dependency packages locally and familiarize yourself with them:
   • pymatgen
   • custodian
   • fireworks
4. Go through the files under settings and overwrite the variables where necessary

Deploy PathFinder HT calculations

1. Prepare an input file in the format like production/Mg_PF_example.txt
2. Run script production/pf_production.py
3. Set up a crontab job to run script like production/crontab.qlaunch.sh regularly at the desired frequency at the calculation clusters.

Deploy ApproxNEB HT calculations

1. Prepare an input file in the format like production/Mg_approx_neb_example.txt, first column being the mp-id, second column being the pair_index defined in the database.
2. Set parameters and run script production/approx_neb_production.py, be noted, it is supposed to run a two-phase calculations, with the first phase calculating just 2 images (image 0 and image 4, the middle image), and arrive at a lower bound for the migration energy, if the migration energy lower bound is smaller than a threshold, then those paths are filtered out for further full MEP calculations. These 2 phases can be switched the adjusting the calculated_images parameter in production/approx_neb_production.py script.
3. Set up a crontab job to run script like production/crontab.qlaunch.sh regularly at the desired frequency at the calculation clusters.

Post calculation

After the HT ApproxNEB calculation is finished, do the following steps:

1. Backup the calculation files. (I do it in Vega)
2. Update the calculation file path in the database after files migrations with script like post_production/db_scripts/update_file_path_vega.py
3. Update the CONTCARs of each calculated image with script like post_production/db_scripts/update_contcar.py

Results Analysis

• After the first-phase HT ApproxNEB calculation, please run post_production/first_tier_filter.py to print out the results.
• After the second-phase HT ApproxNEB calculation, please run post_production/full_mep_results.py to print out the results.