GPAW Installations on Tetralith & Sigma

The GPAW installations are generally maintained by Weine Olovsson (weiol@nsc.liu.se).

"GPAW is a density-functional theory (DFT) Python code based on the projector-augmented wave (PAW) method and the atomic simulation environment (ASE)."

How to run

Note that how to run GPAW might differ between different installations. In some builds it's bundled with Python, ASE etc. and for others they need to be loaded as separate modules.

A sample batch script for running an intel build of GPAW 24.1.0 in parallel:

#!/bin/bash
#SBATCH -A naiss-xxx-yyy
#SBATCH -J jobname
#SBATCH -N 1 --exclusive
#SBATCH -t 2:00:00

export GPAW_SETUP_PATH=/software/sse/manual/GPAW-setups/0.9.20000
module load GPAW/24.1.0-hpc1-intel-2023a-eb

mpprun gpaw python input.py

A sample batch script for running a gcc build of GPAW 24.1.0 in parallel:

#!/bin/bash
#SBATCH -A naiss-xxx-yyy
#SBATCH -J jobname
#SBATCH -N 1 --exclusive
#SBATCH -t 2:00:00

export GPAW_SETUP_PATH=/software/sse/manual/GPAW-setups/0.9.20000

module load Python/3.10.4-env-hpc1-gcc-2022a-eb
module load ASE/3.22.1-hpc1-python
module load GPAW/24.1.0-vdwxc-hpc1-gcc-2022a-eb

mpprun gpaw python input.py

(Note that you should edit the jobname and the account number before submitting.)

Note that mpprun is used for parallel calculations, in this case a full node -N 1 --exclusive gives 32 cores and 32 MPI ranks. GPAW finds the path to the potentials by setting GPAW_SETUP_PATH as shown above.

Available installations

module load GPAW/24.1.0-hpc1-intel-2023a-eb

This build is made with intel and includes MKL, it also includes libxc 6.2.2, but not libvdwxc or ELPA (faster lcao). Depending on type of calculation, it might be faster than gcc builds.

module load Python/3.10.4-env-hpc1-gcc-2022a-eb ASE/3.22.1-hpc1-python
module load GPAW/24.1.0-vdwxc-hpc1-gcc-2022a-eb

This build includes libvdwxc 0.4.0, libxc 6.2.2 and ELPA (which can be used for faster lcao calculations). The corresponding Python and ASE modules need to be loaded together with GPAW.

Old CentOS 7 installations

Several old GPAW modules were retired when the clusters were upgraded to Rocky Linux 9. If there are interest to use the old GPAW versions on Tetralith/Sigma, contact NSC support.