Gromacs []

Questa pagina è in sola lettura. Puoi visualizzare il sorgente, ma non puoi modificarlo. Contatta l'amministratore se pensi che ci sia un errore.
===== Gromacs =====

Several versions of Gromacs are available, check below:

^ Gromacs                                                      ^
| [[http://manual.gromacs.org/archive/4.5/online.html|4.5.7]]  |
| [[http://manual.gromacs.org/documentation/5.1.4|5.1.4]]      |
| [[http://manual.gromacs.org/documentation/5.1.5|5.1.5]]      |
| [[http://manual.gromacs.org/documentation/2016.4|2016.4]]    |
| [[http://manual.gromacs.org/documentation/2016.6|2016.6]]    |
| [[http://manual.gromacs.org/documentation/2018.6|2018.6]]    |
| [[http://manual.gromacs.org/documentation/2018.8|2018.8]]    |
| [[http://manual.gromacs.org/documentation/2019.4|2019.4]]    |
| [[http://manual.gromacs.org/documentation/2019.6|2019.6]]    |
| [[http://manual.gromacs.org/documentation/2020.7|2020.7]]    |
| [[http://manual.gromacs.org/documentation/2021.4|2021.4]]    |
| [[http://manual.gromacs.org/documentation/2021.7|2021.7]]    |
| [[http://manual.gromacs.org/documentation/2022.6|2022.6]]    |

Compilation details:

^ Gromacs  ^ Plumed  ^ Compiler   ^ MPI             ^ CUDA SDK  ^ CUDA RTL  ^ CUDA Architectures  ^
| 4.5.7    |         | GNU 5.4.0  | OpenMPI 1.10.7  |           |           |                     |
| 5.1.4    | 2.3.8   | GNU 5.4.0  | OpenMPI 1.10.7  |           |           |                     |
| 5.1.5    | 2.3.8   | GNU 5.4.0  | OpenMPI 1.10.7  |           |           |                     |
| 2016.4   | 2.3.8   | GNU 5.4.0  | OpenMPI 1.10.7  | 10.0.130  | 10.0.130  | 6.0, 7.0            |
| 2016.6   | 2.3.8   | GNU 5.4.0  | OpenMPI 1.10.7  | 10.0.130  | 10.0.130  | 6.0, 7.0            |
| 2018.6   | 2.4.8   | GNU 5.4.0  | OpenMPI 1.10.7  | 10.0.130  | 10.0.130  | 6.0, 7.0            |
| 2018.6   | 2.4.8   | GNU 7.3.0  | OpenMPI 3.1.4   | 10.0.130  | 10.0.130  | 6.0, 7.0            |
| 2018.8   | 2.4.8   | GNU 5.4.0  | OpenMPI 1.10.7  | 10.0.130  | 10.0.130  | 6.0, 7.0            |
| 2018.8   | 2.4.8   | GNU 7.3.0  | OpenMPI 3.1.4   | 10.0.130  | 10.0.130  | 6.0, 7.0            |
| 2019.4   | 2.8.3   | GNU 5.4.0  | OpenMPI 1.10.7  | 10.0.130  | 10.0.130  | 6.0, 7.0            |
| 2019.4   | 2.8.3   | GNU 7.3.0  | OpenMPI 3.1.4   | 10.0.130  | 10.0.130  | 6.0, 7.0            |
| 2019.6   | 2.8.3   | GNU 5.4.0  | OpenMPI 1.10.7  | 10.0.130  | 10.0.130  | 6.0, 7.0            |
| 2019.6   | 2.8.3   | GNU 7.3.0  | OpenMPI 3.1.4   | 10.0.130  | 10.0.130  | 6.0, 7.0            |
| 2020.7   | 2.8.3   | GNU 8.3.0  | OpenMPI 3.1.6   | 12.0.0    | 12.0.0    | 6.0, 7.0, 8.0       |
| 2021.4   | 2.8.3   | GNU 8.3.0  | OpenMPI 3.1.6   | 12.0.0    | 12.0.0    | 6.0, 7.0, 8.0       |
| 2021.7   | 2.8.3   | GNU 8.3.0  | OpenMPI 3.1.6   | 12.0.0    | 12.0.0    | 6.0, 7.0, 8.0       |
| 2022.6   | 2.8.3   | GNU 8.3.0  | OpenMPI 3.1.6   | 12.0.0    | 12.0.0    | 6.0, 7.0, 8.0       |

For example Gromacs 2022.6 was patched with Plumed 2.8.3 and compiled with the GNU 8.3.0 tool chain and OpenMPI 3.1.6. The GPU enabled version was compiled with CUDA SDK (software development kit) 12.0.0 and is usable with CUDA RTL (run time library) 12.0.0. Code was generated for CUDA architecture 6.0, 7.0, 8.0.

Supported CUDA architectures:

^ GPU               ^ CUDA Architecture ^
| NVIDIA Tesla P100 | 6.0               |
| NVIDIA Tesla V100 | 7.0               |
| NVIDIA Tesla A100 | 8.0               |

Environment modules are available in different flavors:

^ Gromacs  ^ Prerequisites              ^ Flavor  ^ Module      ^
| 4.5.7    | gnu/5.4.0 openmpi/1.10.7   | bdw     | 4.5.7-bdw   |
| 4.5.7    | gnu/5.4.0 openmpi/1.10.7   | bdw     | 4.5.7-cpu   |
| 5.1.4    | gnu/5.4.0 openmpi/1.10.7   | bdw     | 5.1.4-bdw   |
| 5.1.4    | gnu/5.4.0 openmpi/1.10.7   | bdw     | 5.1.4-cpu   |
| 5.1.5    | gnu/5.4.0 openmpi/1.10.7   | bdw     | 5.1.5-bdw   |
| 5.1.5    | gnu/5.4.0 openmpi/1.10.7   | bdw     | 5.1.5-cpu   |
| 2016.4   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2016.4-bdw  |
| 2016.4   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2016.4-cpu  |
| 2016.4   | gnu/5.4.0 openmpi/1.10.7   | gpu     | 2016.4-gpu  |
| 2016.4   | gnu/5.4.0 openmpi/1.10.7   | knl     | 2016.4-knl  |
| 2016.4   | gnu/5.4.0 openmpi/1.10.7   | skl     | 2016.4-skl  |
| 2016.6   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2016.6-bdw  |
| 2016.6   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2016.6-cpu  |
| 2016.6   | gnu/5.4.0 openmpi/1.10.7   | gpu     | 2016.6-gpu  |
| 2016.6   | gnu/5.4.0 openmpi/1.10.7   | knl     | 2016.6-knl  |
| 2016.6   | gnu/5.4.0 openmpi/1.10.7   | skl     | 2016.6-skl  |
| 2018.6   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2018.6-bdw  |
| 2018.6   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2018.6-cpu  |
| 2018.6   | gnu/5.4.0 openmpi/1.10.7   | gpu     | 2018.6-gpu  |
| 2018.6   | gnu/5.4.0 openmpi/1.10.7   | knl     | 2018.6-knl  |
| 2018.6   | gnu/5.4.0 openmpi/1.10.7   | skl     | 2018.6-skl  |
| 2018.6   | gnu7/7.3.0 openmpi3/3.1.4  | bdw     | 2018.6-bdw  |
| 2018.6   | gnu7/7.3.0 openmpi3/3.1.4  | bdw     | 2018.6-cpu  |
| 2018.6   | gnu7/7.3.0 openmpi3/3.1.4  | gpu     | 2018.6-gpu  |
| 2018.6   | gnu7/7.3.0 openmpi3/3.1.4  | knl     | 2018.6-knl  |
| 2018.6   | gnu7/7.3.0 openmpi3/3.1.4  | skl     | 2018.6-skl  |
| 2018.8   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2018.8-bdw  |
| 2018.8   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2018.8-cpu  |
| 2018.8   | gnu/5.4.0 openmpi/1.10.7   | gpu     | 2018.8-gpu  |
| 2018.8   | gnu/5.4.0 openmpi/1.10.7   | knl     | 2018.8-knl  |
| 2018.8   | gnu/5.4.0 openmpi/1.10.7   | skl     | 2018.8-skl  |
| 2018.8   | gnu7/7.3.0 openmpi3/3.1.4  | bdw     | 2018.8-bdw  |
| 2018.8   | gnu7/7.3.0 openmpi3/3.1.4  | bdw     | 2018.8-cpu  |
| 2018.8   | gnu7/7.3.0 openmpi3/3.1.4  | gpu     | 2018.8-gpu  |
| 2018.8   | gnu7/7.3.0 openmpi3/3.1.4  | knl     | 2018.8-knl  |
| 2018.8   | gnu7/7.3.0 openmpi3/3.1.4  | skl     | 2018.8-skl  |
| 2019.4   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2019.4-bdw  |
| 2019.4   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2019.4-cpu  |
| 2019.4   | gnu/5.4.0 openmpi/1.10.7   | gpu     | 2019.4-gpu  |
| 2019.4   | gnu/5.4.0 openmpi/1.10.7   | knl     | 2019.4-knl  |
| 2019.4   | gnu/5.4.0 openmpi/1.10.7   | skl     | 2019.4-skl  |
| 2019.4   | gnu7/7.3.0 openmpi3/3.1.4  | bdw     | 2019.4-bdw  |
| 2019.4   | gnu7/7.3.0 openmpi3/3.1.4  | bdw     | 2019.4-cpu  |
| 2019.4   | gnu7/7.3.0 openmpi3/3.1.4  | gpu     | 2019.4-gpu  |
| 2019.4   | gnu7/7.3.0 openmpi3/3.1.4  | knl     | 2019.4-knl  |
| 2019.4   | gnu7/7.3.0 openmpi3/3.1.4  | skl     | 2019.4-skl  |
| 2019.6   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2019.6-bdw  |
| 2019.6   | gnu/5.4.0 openmpi/1.10.7   | bdw     | 2019.6-cpu  |
| 2019.6   | gnu/5.4.0 openmpi/1.10.7   | gpu     | 2019.6-gpu  |
| 2019.6   | gnu/5.4.0 openmpi/1.10.7   | knl     | 2019.6-knl  |
| 2019.6   | gnu/5.4.0 openmpi/1.10.7   | skl     | 2019.6-skl  |
| 2019.6   | gnu7/7.3.0 openmpi3/3.1.4  | bdw     | 2019.6-bdw  |
| 2019.6   | gnu7/7.3.0 openmpi3/3.1.4  | bdw     | 2019.6-cpu  |
| 2019.6   | gnu7/7.3.0 openmpi3/3.1.4  | gpu     | 2019.6-gpu  |
| 2019.6   | gnu7/7.3.0 openmpi3/3.1.4  | knl     | 2019.6-knl  |
| 2019.6   | gnu7/7.3.0 openmpi3/3.1.4  | skl     | 2019.6-skl  |
| 2020.7   | gnu8/8.3.0 openmpi3/3.1.6  | bdw     | 2020.7-bdw  |
| 2020.7   | gnu8/8.3.0 openmpi3/3.1.6  | bdw     | 2020.7-cpu  |
| 2020.7   | gnu8/8.3.0 openmpi3/3.1.6  | gpu     | 2020.7-gpu  |
| 2020.7   | gnu8/8.3.0 openmpi3/3.1.6  | knl     | 2020.7-knl  |
| 2020.7   | gnu8/8.3.0 openmpi3/3.1.6  | skl     | 2020.7-skl  |
| 2021.4   | gnu8/8.3.0 openmpi3/3.1.6  | bdw     | 2021.4-bdw  |
| 2021.4   | gnu8/8.3.0 openmpi3/3.1.6  | bdw     | 2021.4-cpu  |
| 2021.4   | gnu8/8.3.0 openmpi3/3.1.6  | gpu     | 2021.4-gpu  |
| 2021.4   | gnu8/8.3.0 openmpi3/3.1.6  | knl     | 2021.4-knl  |
| 2021.4   | gnu8/8.3.0 openmpi3/3.1.6  | skl     | 2021.4-skl  |
| 2021.7   | gnu8/8.3.0 openmpi3/3.1.6  | bdw     | 2021.7-bdw  |
| 2021.7   | gnu8/8.3.0 openmpi3/3.1.6  | bdw     | 2021.7-cpu  |
| 2021.7   | gnu8/8.3.0 openmpi3/3.1.6  | gpu     | 2021.7-gpu  |
| 2021.7   | gnu8/8.3.0 openmpi3/3.1.6  | knl     | 2021.7-knl  |
| 2021.7   | gnu8/8.3.0 openmpi3/3.1.6  | skl     | 2021.7-skl  |
| 2022.6   | gnu8/8.3.0 openmpi3/3.1.6  | bdw     | 2022.6-bdw  |
| 2022.6   | gnu8/8.3.0 openmpi3/3.1.6  | bdw     | 2022.6-cpu  |
| 2022.6   | gnu8/8.3.0 openmpi3/3.1.6  | gpu     | 2022.6-gpu  |
| 2022.6   | gnu8/8.3.0 openmpi3/3.1.6  | knl     | 2022.6-knl  |
| 2022.6   | gnu8/8.3.0 openmpi3/3.1.6  | skl     | 2022.6-skl  |

<note>
If required, the Plumed ed CUDA modules are automatically loaded.
</note>

=== GMXLIB environment variable ===

To define the GMXLIB environment variable, add the following lines to the file ''$HOME/.bash_profile'':

<code bash>
GMXLIB=$HOME/gromacs/top

export GMXLIB
</code>

<note>
The path ''$HOME/gromacs/top'' is purely indicative. Modify it according to your preferences.
</note>

==== Job Gromacs OpenMP ==== 

<note>
This will initiate a single process and will result in suboptimal performance.
</note>

=== Gromacs 4.5.7 ===

<note>
Loading the gromacs/4.5.7-bdw or gromacs/4.5.7-cpu modules provides the additional
tools energy2bfac, g_mmpbsa, trj_cavity
</note>

=== Gromacs 5.1.4 ===

<note>
Loading the gromacs/5.1.4-bdw or gromacs/5.1.4-cpu modules provides the additional
tools energy2bfac, g_mmpbsa, trj_cavity
</note>

=== Gromacs 5.1.5 ===

<note>
Loading the gromacs/5.1.5-bdw or gromacs/5.1.5-cpu modules provides the additional
tools energy2bfac, g_mmpbsa, trj_cavity
</note>

Script ''mdrun-omp.sh'' to exclusively request a node and start multiple OpenMP threads (tested with --cpus-per-task=14 and --cpus-per-task=16):

<code bash mdrun-omp.sh>
#!/bin/bash --login
#SBATCH --job-name=mdrun_omp
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=14
#SBATCH --exclusive
#SBATCH --time=0-24:00:00
#SBATCH --mem=120G
#SBATCH --partition=cpu
#SBATCH --qos=cpu
#SBATCH --account=<account>

shopt -q login_shell || exit 1
test -n "$SLURM_NODELIST" || exit 1

module load gnu
module load openmpi
module load gromacs/5.1.5-cpu

# Set OMP_NUM_THREADS to the same value as --cpus-per-task with a fallback in   
# case it isn't set. SLURM_CPUS_PER_TASK is set to the value of --cpus-per-task,
# but only if --cpus-per-task is explicitly set.                                 
export OMP_NUM_THREADS=${SLURM_CPUS_PER_TASK:-1}

gmx mdrun -deffnm topology -pin on
</code>

=== Gromacs 2016.6 ===

Script ''mdrun-omp.sh'' to exclusively request a node and start multiple OpenMP threads (tested with --cpus-per-task=14 and --cpus-per-task=16):

<code bash mdrun-omp.sh>
#!/bin/bash --login
#SBATCH --job-name=mdrun_omp
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=14
#SBATCH --exclusive
#SBATCH --time=0-24:00:00
#SBATCH --mem=120G
#SBATCH --partition=cpu
#SBATCH --qos=cpu
#SBATCH --account=<account>

shopt -q login_shell || exit 1
test -n "$SLURM_NODELIST" || exit 1

module load gnu
module load openmpi
module load gromacs/2016.6-cpu

# Set OMP_NUM_THREADS to the same value as --cpus-per-task with a fallback in   
# case it isn't set. SLURM_CPUS_PER_TASK is set to the value of --cpus-per-task,
# but only if --cpus-per-task is explicitly set.                                 
export OMP_NUM_THREADS=${SLURM_CPUS_PER_TASK:-1}

gmx mdrun -deffnm topology -pin on -plumed plumed.dat
</code>

==== Job Gromacs MPI OpenMP ==== 

<note>
This will initiate multiple MPI processes and will achieve optimal performance.
</note>

=== Gromacs 5.1.5 ===

Script ''mdrun-mpi-omp.sh'' to exclusively request one or more nodes (tested with --nodes=1 and --nodes=2) and start multiple MPI processes (tested with --ntasks-per-node=6 and --ntasks-per-node=8; the number of OpenMP threads will be calculated automatically if --cpus-per-task is not explicitly set):

<code bash mdrun-mpi-omp.sh>
#!/bin/bash --login
#SBATCH --job-name=mdrun_mpi_omp
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=6
#SBATCH --exclusive
#SBATCH --time=0-24:00:00
#SBATCH --mem=120G
#SBATCH --partition=cpu
#SBATCH --qos=cpu
#SBATCH --account=<account>

shopt -q login_shell || exit 1
test -n "$SLURM_NODELIST" || exit 1

module load gnu
module load openmpi
module load gromacs/5.1.5-cpu

# Set OMP_NUM_THREADS to the same value as --cpus-per-task with a fallback in
# case it isn't set. SLURM_CPUS_PER_TASK is set to the value of --cpus-per-task,
# but only if --cpus-per-task is explicitly set.
if [ -n "$SLURM_CPUS_PER_TASK" ]; then
    OMP_NUM_THREADS=$SLURM_CPUS_PER_TASK
else
    if [ -n "$SLURM_NTASKS_PER_NODE" ]; then
        OMP_NUM_THREADS=$((SLURM_CPUS_ON_NODE/SLURM_NTASKS_PER_NODE))
    else
        if [ -n "$SLURM_NTASKS" ] && [ -n "$SLURM_NNODES" ]; then
            OMP_NUM_THREADS=$((SLURM_CPUS_ON_NODE/(SLURM_NTASKS/SLURM_NNODES)))
        else
            OMP_NUM_THREADS=1
        fi
    fi
fi

export OMP_NUM_THREADS

mpirun gmx mdrun -deffnm topology -pin on
</code>

=== Gromacs 2016.6 ===

Script ''mdrun-mpi-omp.sh'' to exclusively request one or more nodes (tested with --nodes=1 and --nodes=2) and start multiple MPI processes (tested with --ntasks-per-node=6 and --ntasks-per-node=8; the number of OpenMP threads will be calculated automatically if --cpus-per-task is not explicitly set):

<code bash mdrun-mpi-omp.sh>
#!/bin/bash --login
#SBATCH --job-name=mdrun_mpi_omp
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=6
#SBATCH --exclusive
#SBATCH --time=0-24:00:00
#SBATCH --mem=120G
#SBATCH --partition=cpu
#SBATCH --qos=cpu
#SBATCH --account=<account>

shopt -q login_shell || exit 1
test -n "$SLURM_NODELIST" || exit 1

module load gnu
module load openmpi
module load gromacs/2016.6-cpu

# Set OMP_NUM_THREADS to the same value as --cpus-per-task with a fallback in
# case it isn't set. SLURM_CPUS_PER_TASK is set to the value of --cpus-per-task,
# but only if --cpus-per-task is explicitly set.
if [ -n "$SLURM_CPUS_PER_TASK" ]; then
    OMP_NUM_THREADS=$SLURM_CPUS_PER_TASK
else
    if [ -n "$SLURM_NTASKS_PER_NODE" ]; then
        OMP_NUM_THREADS=$((SLURM_CPUS_ON_NODE/SLURM_NTASKS_PER_NODE))
    else
        if [ -n "$SLURM_NTASKS" ] && [ -n "$SLURM_NNODES" ]; then
            OMP_NUM_THREADS=$((SLURM_CPUS_ON_NODE/(SLURM_NTASKS/SLURM_NNODES)))
        else
            OMP_NUM_THREADS=1
        fi
    fi
fi

export OMP_NUM_THREADS

mpirun gmx mdrun -deffnm topology -pin on -plumed plumed.dat
</code>

<note>
Plumed module is automatically loaded.
</note>

=== Gromacs 2019.6 ===

Script ''mdrun-mpi-omp-9x3.sh'' to request one node and start 9 MPI processes and 3 OpenMP threads for each MPI process (27 threads overall):

<code bash mdrun-mpi-omp-9x3.sh>
#!/bin/bash --login
#SBATCH --job-name=mdrun_mpi_omp_9x3
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=28
#SBATCH --time=0-24:00:00
#SBATCH --mem=80G
#SBATCH --partition=cpu
#SBATCH --qos=cpu
#SBATCH --account=<account>

shopt -q login_shell || exit 1
test -n "$SLURM_NODELIST" || exit 1

module load gnu7
module load openmpi3
module load gromacs/2019.6-cpu

module list

OMP_NUM_THREADS=3

export OMP_NUM_THREADS

mpirun -np 9 gmx mdrun -deffnm meta -pin on -plumed plumed.dat
</code>

<note>
Plumed module is automatically loaded.
</note>

=== Gromacs 2021.7 ===

Script ''mdrun-mpi-omp-9x3.sh'' to request one node and start 9 MPI processes and 3 OpenMP threads for each MPI process (27 threads overall):

<code bash mdrun-mpi-omp-9x3.sh>
#!/bin/bash --login
#SBATCH --job-name=mdrun_mpi_omp_9x3
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=28
#SBATCH --time=0-24:00:00
#SBATCH --mem=80G
#SBATCH --partition=cpu
#SBATCH --qos=cpu
#SBATCH --account=<account>

shopt -q login_shell || exit 1
test -n "$SLURM_NODELIST" || exit 1

module load gnu8
module load openmpi3
module load gromacs/2021.7-cpu

module list

OMP_NUM_THREADS=3

export OMP_NUM_THREADS

mpirun -np 9 gmx mdrun -deffnm meta -pin on -plumed plumed.dat
</code>

====  Job Gromacs OpenMP GPU ==== 

=== Gromacs 2016.6 ===

The ''mdrun-omp-cuda.sh'' script runs a 4 core with 1 CUDA GPU Gromacs job:

<code bash mdrun-omp-cuda.sh>
#!/bin/bash --login
#SBATCH --job-name=mdrun_omp_cuda
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=4
#SBATCH --gres=gpu:1
#SBATCH --time=0-24:00:00
#SBATCH --mem=16G
#SBATCH --partition=gpu
#SBATCH --qos=gpu
#SBATCH --account=<account>

shopt -q login_shell || exit 1
test -n "$SLURM_NODELIST" || exit 1

module load gnu
module load openmpi
module load gromacs/2016.6-gpu

# Set OMP_NUM_THREADS to the same value as --cpus-per-task with a fallback in
# case it isn't set. SLURM_CPUS_PER_TASK is set to the value of --cpus-per-task,
# but only if --cpus-per-task is explicitly set.
export OMP_NUM_THREADS=${SLURM_CPUS_PER_TASK:-1}

gmx mdrun -deffnm topology -pin on -dlb auto -plumed plumed.dat
</code>

====  Job Gromacs MPI GPU ==== 

=== Gromacs 2019.6 ===

Script ''mdrun-mpi-omp-gpu-4x2.sh'' to request one node and start 4 MPI processes and 2 OpenMP threads for each MPI process (8 threads overall) with 2 CUDA GPU's:

<code bash mdrun-mpi-omp-gpu-4x2.sh>
#!/bin/bash --login
#SBATCH --job-name=mdrun_mpi_omp_gpu
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=8
#SBATCH --gres=gpu:2
#SBATCH --time=0-24:00:00
#SBATCH --mem=16G
#SBATCH --partition=gpu
#SBATCH --qos=gpu
#SBATCH --account=<account>

shopt -q login_shell || exit 1
test -n "$SLURM_NODELIST" || exit 1

module load gnu7
module load openmpi3
module load gromacs/2019.6-gpu

OMP_NUM_THREADS=2

export OMP_NUM_THREADS

mpirun -np 4 gmx mdrun -deffnm meta -dlb auto -plumed plumed.dat
</code>

=== Gromacs 2021.7 ===

Script ''mdrun-mpi-omp-gpu-4x2.sh'' to request one node and start 4 MPI processes and 2 OpenMP threads for each MPI process (8 threads overall) with 2 CUDA GPU's:

<code bash mdrun-mpi-omp-gpu-4x2.sh>
#!/bin/bash --login
#SBATCH --job-name=mdrun_mpi_omp_gpu
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=8
#SBATCH --gres=gpu:2
#SBATCH --time=0-24:00:00
#SBATCH --mem=16G
#SBATCH --partition=gpu
#SBATCH --qos=gpu
#SBATCH --account=<account>

shopt -q login_shell || exit 1
test -n "$SLURM_NODELIST" || exit 1

module load gnu8
module load openmpi3
module load gromacs/2021.7-gpu

OMP_NUM_THREADS=2

export OMP_NUM_THREADS

mpirun -np 4 gmx mdrun -deffnm meta -dlb auto -plumed plumed.dat
</code>