Application Guide - Gaussian


Gaussian is a commercial computational chemistry software for modelling electronic structures.

Note: While the use of Gaussian is free for University researchers, the ARC team has to enable users to run Gaussian on an individual basis; if you are planning to run Gaussian, let the ARC team know.

The guide shows how to

  • prepare a Gaussian job submission script and
  • submit and run a Gaussian job.


    Gaussian on the ARC systems

    Gaussian 03 , Gaussian 09 and Gaussian 16 are available on all the ARC systems and were built from source.  They can be loaded using the commands module load gaussian03 and module load gaussian16 respectively.  (There are several modules available for each in fact but users should normally use the default module.)

    Gaussian is a multi-threaded application and users are advised to take advantage of this feature as it leads to a faster execution.  Running Gaussian multi-threaded is controlled from the header of the Gaussian com input files.


    Running a Gaussian job

    Here is an example of how to run the valinomycin test (part of the Gaussian distribution) on a cluster node.

    First, load the Gaussian 09 module

    module load gaussian09


    Then, edit the file to set the number of threads.  This should be 16 on the clusters in order to match the availability of cores per compute node. 



    Then, a slurm submission script to run this test on a cluster node is (Torque variant on the left, Slurm variant on the right):


    #SBATCH --nodes=1
    #SBATCH --time=01:00:00
    #SBATCH --job-name=testGaussian


    g09 < > test397.log


    The script requires 1 node, launches Gaussian 09 taking the input from the file and forces the standard output to the file test397.log.

    To launch into execution, issue the command (Torque sumission on the left, Slurm submission on the right)





    The Gaussview software is also installed on the ARC clusters for convenience. However, we recommend that the desktop version of Gaussview is used to prepare input for Gaussian, as running graphical applications across the network is less efficient. The Department of Chemistry has a site licence for Gaussview and can be contacted for details at the following email address: