GROMACS is a versatile package to perform molecular dynamics, i.e. simulate the Newtonian equations of motion for systems with hundreds to millions of particles.
This is a GPU enabled build, containing both MPI and threadMPI binaries.
Available Modules¶
module load GROMACS/2021.6-gimkl-2020a-cuda-11.6.2-hybrid-PLUMED-2.8.0GROMACS (proper name, Not an acronym) is a versatile package to perform molecular dynamics, i.e. simulate the Newtonian equations of motion for systems with hundreds to millions of particles.
It is primarily designed for biochemical molecules like proteins, lipids and nucleic acids that have a lot of complicated bonded interactions, but since GROMACS is extremely fast at calculating the nonbonded interactions (that usually dominate simulations) many groups are also using it for research on non-biological systems, e.g. polymers.
GROMACS is available to anyone at no cost under the terms of the GNU Lesser General Public Licence. Gromacs is a joint effort, with contributions from developers around the world: users agree to acknowledge use of GROMACS in any reports or publications of results obtained with the Software.
Examples¶
For when only one CPU is required, generally as part of a job array
#!/bin/bash -e
#SBATCH --job-name GROMACS-serial
#SBATCH --time 00:05:00 # Walltime
#SBATCH --account nesi99991 # Your project ID
#SBATCH --mem 1500 # How much memory.
module load GROMACS/2021.6-gimkl-2020a-cuda-11.6.2-hybrid-PLUMED-2.8.0
# Note: In version 2021.5 and older use `gmx-serial` instead of `gmx`
srun gmx mdrun -s input.tpr -o trajectory.trr -c struct.gro -e energies.edr
Uses a nodes shared memory for communication.
#!/bin/bash -e
#SBATCH --job-name GROMACS-shared-mem
#SBATCH --time 00:05:00 # Walltime
#SBATCH --account nesi99991 # Your project ID
#SBATCH --cpus-per-task 8 # Will use 8 CPUs
#SBATCH --mem 1500 # How much memory.
module load GROMACS/2021.6-gimkl-2020a-cuda-11.6.2-hybrid-PLUMED-2.8.0
# Note: In version 2021.5 and older use `gmx-serial` instead of `gmx`
srun gmx mdrun -ntomp ${SLURM_CPUS_PER_TASK} -s input.tpr -o trajectory.trr -c struct.gro -e energies.edr
Should only be used in the case you need more CPUs than available on a single node.
```sl
!/bin/bash -e¶
SBATCH --job-name GROMACS-multi-node¶
SBATCH --time 00:05:00 # Walltime¶
SBATCH --account nesi99991 # Your project ID¶
SBATCH --nodes 2 # wiLL use 2 nodes.¶
SBATCH --mem 1500 # How much memory.¶
module load GROMACS/2021.6-gimkl-2020a-cuda-11.6.2-hybrid-PLUMED-2.8.0 srun gmx-mpi mdrun-mpi -ntomp ${SLURM_CPUS_PER_TASK} -nomp ${SLURM_NNODES) -s input.tpr -o trajectory.trr -c struct.gro -e energies.edr
For more information on using GPUs see GPU use on NeSI
#!/bin/bash -e
#SBATCH --job-name GROMACS-multi-node
#SBATCH --time 00:05:00 # Walltime
#SBATCH --account nesi99991 # Your project ID
#SBATCH --gpus-per-node 1
#SBATCH --cpus-per-task 8
#SBATCH --mem 1500 # How much memory.
module load GROMACS/2021.6-gimkl-2020a-cuda-11.6.2-hybrid-PLUMED-2.8.0
# Note: In version 2021.5 and older use `gmx-serial` instead of `gmx`
srun gmx mdrun -ntomp ${SLURM_CPUS_PER_TASK} -s input.tpr -o trajectory.trr -c struct.gro -e energies.edr
Performance¶
GROMACS performance depends on several factors, such as usage (or lack thereof) of GPUs, the number of MPI tasks and OpenMP threads, the load balancing algorithm, the ratio between the number of Particle-Particle (PP) ranks and Particle-Mesh-Ewald (PME) ranks, the type of simulation being performed, force field used and of course the simulated system. For a complete set of GROMACS options, please refer to GROMACS documentation.
Each GROMACS environment module contains two executables one built with shared memory parallelism gmx,
and one with distrubuted memorory parallelism (MPI) gmx-mpi. which can run across multiple nodes,
ie: with --ntasks > 1.
Warning
In versions of GROMACS older than GROMACS/2025.2-foss-2023a-cuda-12.5.0-hybrid
the gmx executable is instead called gmx-serial.
CUDA¶
GROMACS is built with CUDA support, but that is optional to use - it will run without a GPU.
MPI¶
If you do elect to use gmx-mpi, note that hybrid parallelisation (i.e. with --cpus-per-task > 1) can be
more efficient than MPI-only parallelisation. With hybrid parallelisation, it is important to run
mdrun_mpi with the -ntomp <number> option, where <number> should
be the number of CPUs per task. You can make sure the value is correct
by using -ntomp ${SLURM_CPUS_PER_TASK}.
Checkpointing¶
The -cpt 30 option instructs Gromacs to
write a full checkpoint file every 30 minutes.
We reccomend including this flag for long running jobs.
You can restart from a
checkpoint file using the -cpi flag, thus: -cpi state.cpt.