Fleur#
Fleur is a tool for calculating material properties with density functional theory (DFT) and related methods. It is released under the MIT licence. For further information, including tutorials, please see the official documentation
Available versions on Aire#
Version |
Load Command |
|---|---|
MaX8.0 (no HDF5 support) |
|
MaX8.0 (with HDF5 support) |
|
Note that there are two versions available on Aire; with and without support for HDF5. For more intensive calculations, it is likely that the HDF5 version will be faster, with the opposite true for simpler runs.
How to run a job#
A fleur job generally runs in two steps. First, an .xml-format input file is generated from a human readable text file, using the input generator. This can be performed on a login node, as it is not computationally intensive. Please see the tutorial for further information.
[user@login1[aire] my_folder]$ module load fleur
Loading fleur/max8_hdf5/openmpi-5.0.5-gcc-14.2.0_lapack-3.12.0
Loading requirement: gcc/14.2.0 lapack/3.12.0 openmpi/5.0.5/gcc-14.2.0
[user@login1[aire] my_folder]$ inpgen -f my_input.txt
Welcome to FLEUR - inpgen (www.flapw.de)
MaX-Release 8.0 (www.max-centre.eu)
...
Run finished successfully
Stop message:
All done
...
[user@login1[aire] my_folder]$
Fleur supports distributed-memory parallelism with OpenMPI. To run a Fleur job, call Fleur from the same directory as the generated input file. Below is an example job submission script.
#!/bin/bash
#SBATCH --job-name=fleur_example
#SBATCH --time=00:20:00 # Request running time
#SBATCH --mem=10G # Request memory
#SBATCH --nodes=2 # Number of nodes
#SBATCH --ntasks-per-node=3 # Number of tasks per node
#SBATCH --cpus-per-task=1 # Number of CPUs per task
# Load Fleur module and dependencies
module load fleur/max8_hdf5
# Run the job using MPIrun, passing in number of tasks
mpirun -n $SLURM_NTASKS_PER_NODE fleur_MPI