GPAW

GPAW is a density-functional theory (DFT) and beyond code based on the projector-augmented wave (PAW) method and the atomic simulation environment (ASE). The wave functions can be described with plane waves, uniform real-space grids, and atom-centered basis functions.

Some features of the software include:

total energy calculations
structural optimizations
different boundary conditions (finite, wire, slab, bulk)
efficient parallelization
excited state properties within time-dependent density-functional theory

Available

Roihu

Roihu-CPU: 25.7.0
Check all available versions (and default version) with module avail gpaw
The installation includes the following libraries: MPI, OpenMP, ScaLAPACK, ELPA, FFTW, libxc, libvdwxc, DFT-D3, DFT-D4
See GPAW documentation on parallel runs for instructions on how to enable high-performance libraries in the input script
The PAW setups are installed through gpaw_data package
Use gpaw info for detailed version information

License

GPAW is free software available under GPL, version 3+.

Usage

Since the default version, that is available with module load gpaw, is subject to change when new versions are installed, we recommend to always load a specific GPAW version:

module load gpaw/version

Note

GPAW calculations are run with the gpaw python command.

Batch script examples

Roihu-CPU (partial node)Roihu-CPU (full nodes)

#!/bin/bash
#SBATCH --job-name=gpaw
#SBATCH --account=<project>
#SBATCH --partition=small
#SBATCH --time=00:30:00
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=2
#SBATCH --cpus-per-task=1
#SBATCH --mem-per-cpu=1000M

# Set the number of threads based on cpus-per-task
export OMP_NUM_THREADS=${SLURM_CPUS_PER_TASK:-1}

# Place and bind threads to single cores
# Comment the following lines if binding is not desired
export OMP_PLACES=cores
export OMP_PROC_BIND=spread

# Run GPAW
srun gpaw python input.py

#!/bin/bash
#SBATCH --job-name=gpaw
#SBATCH --account=<project>
#SBATCH --partition=medium
##SBATCH --partition=large  # uncomment if using 6 or more nodes
#SBATCH --time=00:30:00
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=192 --cpus-per-task=2  # The product should be 384

# Test different values of
# --ntasks-per-node and --cpus-per-task above
# for your use case and use the values that give the best performance

# Set the number of threads based on cpus-per-task
export OMP_NUM_THREADS=${SLURM_CPUS_PER_TASK:-1}

# Place and bind threads to single cores
# Comment the following lines if binding is not desired
export OMP_PLACES=cores
export OMP_PROC_BIND=spread

# Run GPAW
srun gpaw python input.py

More information

GPAW home page