Namd¶

Support tier: 3

Read information about support tiers.

Installed versions¶

Resource	Version
Arrhenius-GPU/cpe25.09	3.0.2

Read information about how to load this software in your environment by searching for Lmod module.

General information¶

NAMD is a parallel molecular dynamics code designed for high-performance simulation of large biomolecular systems. For more information see the NAMD homepage.

NAMD is available to users as a module - we have a non-exclusive, non-commercial use license for academic purposes.

Citations

See the Acknowledge the TCB Group page for specifics on how to acknowledge the use of NAMD.

You can also register at the TCB Group to use NAMD, even when you do not download it.

Dardel¶

How to use¶

Example job script for a regular NAMD run on 2 nodes:

#!/bin/bash
#SBATCH -A XXXX-XX-XX
#SBATCH -J namdjob
#SBATCH -t 00:10:00
#SBATCH --nodes=2
#SBATCH -p main
#SBATCH --ntasks-per-node=128
# load the NAMD module
ml PDC/<version>
ml NAMD/3.0b
# Run namd
srun namd2 input.namd > output_file

Example job script for a 2-node shared-memory NAMD run with one MPI process per node:

#!/bin/bash
#SBATCH -A XXXX-XX-XX
#SBATCH -J namdjob
#SBATCH -t 00:10:00
#SBATCH --nodes=2
#SBATCH -p main
#SBATCH --ntasks-per-node=1
# load the NAMD module
ml PDC/<version>
ml NAMD/3.0b
# Run namd
srun -n 2 namd2 +ppn 31 input.namd > output_file

Example job script for a 2-node shared-memory NAMD run with 4 MPI processes per node and explicit mapping of communication and processing cores:

#!/bin/bash
#SBATCH -A XXXX-XX-XX
#SBATCH -J namdjob
#SBATCH -t 00:10:00
#SBATCH --nodes=2
#SBATCH -p main
#SBATCH --ntasks-per-node=4
# load the NAMD module
ml PDC/<version>
ml NAMD/2.14
# Run namd
srun -n 8 namd2 +ppn 7 +pemap 1-7,9-15,17-23,25-31 +commap 0,8,16,24 input.namd > output_file

Arrhenius¶

How to use¶

Example job script for a GPU NAMD run on 1 GPU card:

#!/bin/bash
#SBATCH -A XXXX-XX-XX
#SBATCH -J namdjob
#SBATCH -t 00:10:00
#SBATCH -p gpu
#SBATCH --gpus=1
# load the NAMD module
ml GPU/NAMD/3.0.2-gcccuda-2026.03-cu13.0-es
# Run namd
namd3 input.namd +p1 +setcpuaffinity +devices 0 > output_file

A good performance can be achieved if the number of processes (1 in this example p1) matches the number of available GPUs (also 1 in the present case).

Check if the your simulation type supports the GPU resident mode as the performace of the simulation could increase by a 10x factor depending on the simulation size.