Gromacs is a very efficient engine to perform molecular dynamics simulations and energy minimization particularly for proteins. However, it can also be used to model polymers, membranes and e.g. coarse grained systems. It also comes with plenty of analysis scripts.


Version on CSC's Servers

Sisu,Taito: 4.6.7, 5.1.5, 2016.x, 2018.5

Taito-GPU: 5.1.5, 2016.5, 2018.3 (NOTE: Don't use GPU versions 2018.1 or 2018.2, see link)

Check available versions with module avail gromacs

The 5.1.5 (Taito) and 2016.3/4, 2018.5 (non-GPU) versions also include plumed

Note, Gromacs versions with the hsw (Haswell) tag won't run on the login node, but give better performance on Haswell compute nodes. For preparing jobs (gmx grompp, etc.) use the snb modules on the login node, but the appropriate version in your batch script. The same applies also for GPU enabled versions. Some modules have only mdrun_mpi for parallel runs. See output of module command whether to use gmx_mpi mdrun or mdrun_mpi.


To start working with Gromacs in Sisu use the module command:

module load gromacs/version

In Taito use

module load gromacs-env/version

Omitting version will initialise the default version which will change when bugs are fixed or new versions are accepted stable. Version 2016.5 is currently the default.

Gromacs is also available on the Finnish Grid Infrastructure (FGI).  Please see below for the example on how to run Gromacs on the FGI.


The batch submission is different on all machines. Please consult the server manuals or the examples at the bottom of the page.
Examples of Gromacs 4 jobs on the FGI.

Other versions

You can check for other versions with the command module avail or module spider on Taito

After initialisation you can check, e.g. with

which gmx

which shows where the code lives and hints how to modify the batch script initialisation.

Which version to use?

We recommend using the latest versions as they have most bugs fixed and tend to be faster. If you switch the major version, check that the results are comparable. See below for a brief performance test using the d.dppc system (~100k atoms, PME). Note, that the 2018.1 version is more than 50% faster than 4.6.7 and likely scales better to large number of cores. However, remember always to test scaling for your own systems to select the right number of cores to use.

  (Node archictecture / Code optimization to) cores gmx 4.5.6 gmx 4.6.7 gmx 5.1.5 gmx 2016.5 gmx 2018.1
Taito Hsw/Hsw 48 - - 45.505 48.407 49.369
Taito Hsw/Snb 48 21.523 21.9 38.953 40.987 41.318
Taito Snb/Snb 48 20.694 20.513 28.559 34.958 39.906
Sisu Hsw/Hsw 48 - 30.104 39.935 48.979 49.624
Sisu Hsw/Hsw 96 - 62.499 80.769 86.638 96.194
Taito-gpu k80:1/Hsw 6 - - - 23.964 27.18
Taito-gpu p100:1/Hsw 7 - - - 31.703 55.041
Taito-gpu p100:4/Hsw 28 - - - - 62.868

The second column indicates which Node architecture was used (first part) and which code optimizations were used (second). For example, the last line means that four P100 GPGPU cards were used (and on the next column that for each 7 CPU cores were allocated). The performance numbers are in ns/day i.e. more is better.

From the table you can see, that or this system 1) the Haswell optimized Gromacs on Haswell Nodes is 15-20% faster than Sandby Bridge optimized Gromacs on  Haswell node, 2) on Taito Sandy bridge nodes version 2018.1 is twice as fast as 4.6.7, 3) it does not make sense to use 4 GPGPU-cards as the speedup compared to using one is only 1.14 fold (while it should be close to 4).

Visualising trajectories

In addition to ngmx program in Gromacs, trajectory files can be visualized with the following programs:

  • PyMOL molecular modeling system.
  • VMD visualizing program for large biomolecular systems. For remote visualization consider NoMachine and TurboVNC + GPUPU

Examples and links




Please cite the following papers:

  • GROMACS 4: Algorithms for Highly Efficient, Load-Balanced, and Scalable Molecular Simulation. Hess, B., Kutzner, C., van der Spoel, D. and Lindahl, E. J. Chem. Theory Comput., 4, 435-447 (2008).
  • GROMACS: Fast, Flexible and Free. D. van der Spoel, E. Lindahl, B. Hess, G. Groenhof, A. E. Mark and H. J. C.Berendsen, J. Comp. Chem. 26 (2005) pp. 1701-1719