UQ RCC HPC Documentation

High Performance Computing guides and policies for the University of Queensland.

View the Project on GitHub UQ-RCC/hpc-docs

R Versions on Bunya

TO BE COMPLETED

David Green 07/11/2025

What is R?

R is a programming language that developed out of statistical computing and data analysis. It has been enhanced with thousands of packages for a wide variety of computational work. It continues to be a popular choice amongst Bunya users for simulations and analysis.

Why so many versions on Bunya?

Refer to the Bunya User Guide section about software modules.

Older Bare Metal Versions

When Bunya was first launched we provided R built using our centralised build system. The following versions included the core R system as well as a number of useful additional packages.

r/4.1.0-foss-2021a
r/4.2.1-foss-2021a
r/4.2.1-foss-2022a

Newer Bare Metal Versions

The build mechanism changed for R 4.3 and R 4.4.0. The additional packages (numbering approximately 1,200) were built as separate items.

r/4.3.3-gfbf-2023a
r/4.4.0-gfbf-2023a
r/4.4.0-combo-EPYC3-only
r/4.4.2-gfbf-2024a

Unfortunately, the r/4.4.0-gfbf-2023a installation was incomplete. The 1,200 or so additional packages did not build for EPYC4 CPUs. If you do choose to use this version, then you may need to build a great many packages for yourself.
I recommend that you avoid using r/4.4.0-gfbf-2023a.

The r/4.4.2-gfbf-2024a is a new build after the Rocky 9 operating system upgrade.

If you target the older EPYC3 CPU nodes, you can use the purpose built r/4.4.0-combo-EPYC3-only

Software container versions

Please refer to the Basics.apptainer for information about software containers. Software containers do not run on login nodes.

The R programming language underpins the integrated development environment (IDE) called RStudio.
Since the launch of onBunya, the RStudio software (and R within) are provided on Bunya as software containers. There are several versions available.

If you need to use a software container version of R without RStudio that is OK. The following software modules use the RStudio containers to provide access to the R version within.

Module Name Features
r/4.4.1 Software container with approximately 1,200 popular R packages from CRAN.
r/4.4.2 Software container with approximately 1,200 popular R packages from CRAN.
r/4.4.2-heavy Software container with approximately 25,000 R packages from CRAN.</br>Requires >24GB of RAM to load.
r/4.5.1 Software container with approximately 1,200 popular R packages from CRAN.

[!CAUTION] If you use a bare metal R 4.4.x version and switch to a software container 4.4.x version, then you should delete your personal library space for that version e.g. $HOME/R/x86_64-pc-linux-gnu-library/4.4/ and rebuild all your personal modules.

How to access R?

Using R within RStudio in onBunya

The RStudio software is available via the applications menu in the onBunya graphical desktop service.

Using R within a command terminal in onBunya

Launch a terminal within onBunya.
Use a module load of the appropriate R version you require.
The commands, R and Rscript, will be automatically found in your path.

Using R in an interactive batch job

Use a module load of the appropriate R version you require.
The commands, R and Rscript, will be automatically found in your path.

Using R in a regular batch job

Use a module load of the appropriate R version you require in your job script.
The commands, R and Rscript, will be automatically found in your path.

Problems with accessing /QRISdata from container versions of R.

Yes this is an issue. When you launch a software container you do so as your default group which won’t be allowed to access your RDM data.
Pre-fetch your data from /QRISdata onto $TMPDIR or /scratch.