Advice for New Students on Getting Started With DFT

To install VASP on your own computer, use the provided VASP install directions with the install files James provides. You are encouraged to explore and modify the VASP source code for your own needs. VASP should also already be installed on any supercomputer the group has an allocation on.

You can find a set of simple tutorials in this tape archive file VASP Tutorial.tar

However, for understanding typical calculations that the group does, it is best to look at VASP files from other group members. Josh’s INCAR contains a good sampling of tags with usual parameters the group uses. You should also look at the group GitHub to see more sample VASP files and job scripts.

The group has several computational resources you can utilize. Quest is Northwestern’s high performance computing (HPC) cluster. The group has several nodes reserved for our own use and there are no limits on computer hours provided you use the “buyin” queue in your job file. However, Quest jobs may wait longer in queue and take longer to run than on a larger cluster such as Texas Advanced Computing Center’s (TACC) Stampede or Argonne National Lab’s Carbon. We have hours allotted on both of those clusters that someone should be keeping track of at all times. Generally, we shift jobs between clusters whenever utilization is high. Professor Rondinelli is in control of all these accounts and will be the one to get you registered. It is possible to have VASP compiled on your personal desktop for running simple jobs, but the majority of jobs you run will likely be on a cluster.

On Quest, the job submission command is ‘msub’. On Stampede, it is ‘sbatch’. You can see more commands and settings by viewing the job scheduler documentation for MOAB (Quest) and SLURM (Stampede). Sample job submission scripts are available on the group GitHub.