Quantum ESPRESSO¶
Quantum ESPRESSO is a suite for first-principles electronic-structure calculations and materials modeling, distributed for free and as free software under the GNU General Public License. It is based on density-functional theory, plane wave basis sets, and pseudopotentials.
Quantum ESPRESSO module¶
Quantum ESPRESSO has to be loaded using Lmod prior to running it.
$ module load QuantumESPRESSO
Loading the module for QuantumESPRESSO will also give you access to yambo binary and some utilities such as wannier.\
You can also list all available versions using Lmod's spider command:
$ module spider QuantumESPRESSO
Software version¶
Here you can check the available versions for QuantumESPRESSO in the different clusters
QuantumESPRESSO/old_6.5-intel-2019b
QuantumESPRESSO/4.3.2-intel-2017a
QuantumESPRESSO/5.3-intel-2017b
QuantumESPRESSO/6.2-intel-2017b-hybrid
QuantumESPRESSO/6.2-intel-2017b-mt
QuantumESPRESSO/6.2-intel-2017b
QuantumESPRESSO/6.2.1-develop-54a55df0-intel-2017b-hybrid
QuantumESPRESSO/6.2.1-develop-54a55df0-intel-2017b-mt
QuantumESPRESSO/6.2.1-develop-54a55df0-intel-2017b
QuantumESPRESSO/6.2.1-intel-2017b-hybrid
QuantumESPRESSO/6.2.1-intel-2017b-mt
QuantumESPRESSO/6.2.1-intel-2017b
QuantumESPRESSO/6.3-intel-2018a
QuantumESPRESSO/6.4-intel-2018a
QuantumESPRESSO/6.5-intel-2019b
QuantumESPRESSO/6.7-intel-2019b-thermo
QuantumESPRESSO/6.7-intel-2019b
QuantumESPRESSO/6.5-intel-2019b
QuantumESPRESSO/6.6-intel-2019b
QuantumESPRESSO/6.6-intel-2020a
QuantumESPRESSO/6.8-intel-2021a
QuantumESPRESSO/7.2-foss-2023a
QuantumESPRESSO/6.7-NVHPC-23.1-CUDA-12.1.1-serial
QuantumESPRESSO/7.1-foss-2022a
QuantumESPRESSO/7.1-intel-2022a
QuantumESPRESSO/7.2-foss-2023a
QuantumESPRESSO/7.2-NVHPC-23.1-CUDA-12.1.1
How to run Quantum ESPRESSO on Atlas¶
Batch script for parallel execution of Quantum ESPRESSO on Atlas EDR
#!/bin/bash
#SBATCH --qos=regular
#SBATCH --job-name=QuantumESPRESSO_job
#SBATCH --mem=200gb
#SBATCH --cpus-per-task=1
#SBATCH --nodes=8
#SBATCH --ntasks-per-node=48
#SBATCH --output=%x-%j.out
#SBATCH --error=%x-%j.err
module load QuantumESPRESSO/6.6-intel-2020a
srun --cpu_bind=cores pw.x -input input.in
Warning
After loading the module the value of OMP_NUM_THREADS is equal to 1. If you want to activate the hybrid MPI+OpenMP parallelism you would need to adjust the value of said variable and tweak the batch script accordingly.
Batch script for MPI+OpenMP execution of Quantum ESPRESSO on Atlas EDR
#!/bin/bash
#SBATCH --qos=regular
#SBATCH --job-name=QuantumESPRESS_job
#SBATCH --mem=200gb
#SBATCH --cpus-per-task=8
#SBATCH --nodes=8
#SBATCH --ntasks-per-node=6
#SBATCH --output=%x-%j.out
#SBATCH --error=%x-%j.err
module load QuantumESPRESSO/6.6-intel-2020a
export OMP_NUM_THREADS=$SLURM_CPUS_PER_TASK
srun --cpu_bind=cores pw.x -input input.in
Batch script for GPU execution of Quantum ESPRESSO on Atlas EDR
#!/bin/bash
#SBATCH --qos=regular
#SBATCH --job-name=QuantumESPRESS_job
#SBATCH --mem=40gb
#SBATCH --cpus-per-task=2
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=4
#SBATCH --gpu=rtx3090:2
#SBATCH --output=%x-%j.out
#SBATCH --error=%x-%j.err
module load QuantumESPRESSO/7.2-NVHPC-23.1-CUDA-12.1.1
mpirun pw.x -input input.in >> qe.out