The Gurobi mathmatical optimisation solver is available to use on HPC. You can find the official documentation on their website.

Gurobi module

To access the Gurobi solver you need to load the gurobi module. You can do this with the command:

$ module add test gurobi


This module is new and currently being tested so can only be loaded by loading the test module first.

This sets a number of environment variables which ensure any Gurobi APIs know how to communicate with the solver, these include:


Python setup

To use Gurobi in a python script you will need to install the Gurobipy package. You can do this using either pip or conda with the following commands:

# installing with pip
$ pip install --user -i gurobipy

# installing with conda
$ conda install -c gurobi gurobi

You can then use the Gurobi Python API by adding the line import gurobi to your python script and calling functions from the module.

R setup


The Gurobi R package is currently only available for R version 4.0.3

To use Gurobi with the R language you need to install the R gurobi package and the R package slam. This is not on CRAN so to install this on ARC3/ARC4 you must install the package using one of the following methods:

# install Gurobi and slam within the R console
> install.packages(c(Sys.getenv("GUROBI_R_PKG"),"slam"))
# install Gurobi and slam from the shell
$ R -e "install.packages(c('$GUROBI_R_PKG','slam'))"

Once you have installed the package you can then load it within the R environment using the line library(gurobi). Once the package is loaded you can use Gurobi R API functions to use the solver.

Installing R 4.0.3 via Anaconda module

You can install R 4.0.3 on the HPC by using the Anaconda module via the conda package manager. In its simplest form you can do this with the following commands:

# add the anaconda module
$ module add anaconda

# create a conda environment in which we install R 4.0.3 from the conda forge channel
$ conda create -n R403 -c conda-forge r-base=4.0.3

# to enter the environment and use R 4.0.3
$ source activate R403

Matlab setup

To configure MATLAB to use the Gurobi module you will need to have loaded the Gurobi module and within the MATLAB command window run the following commands:

% change directory into the Gurobi matlab directory

% run the gurobi_setup.m file to configure MATLAB

This adds the Gurobi MATLAB directory to your MATLAB path in your current MATLAB session. To permanently save this directory to your MATLAB path you can use the following commands after performing the above:

% save your current (updated) MATLAB path in your home directory

Once you have run gurobi_setup you should now have access to the Gurobi MATLAB API and can run Gurobi specific functions.