BwUniCluster3.0/Software/R
|
The main documentation is available on the cluster via |
| Description | Content |
|---|---|
| module load | math/R |
| License | GPL |
| Citing | n/a |
| Links | Homepage | Documentation |
| Graphical Interface | No |
| Plugins | User dependent |
Description
R is a free and open-source statistical programming environment based on S, originally developed by John Chambers and colleagues at Bell Laboratories (formerly AT&T, now Lucent Technologies) in the 1970s. R can be considered a modern implementation of S and is known to be highly backward-compatible, meaning that code developed in a previous version will still run in a new version.
R provides a wide variety of statistical (linear and nonlinear modelling, classical statistical tests, time-series analysis, classification, clustering, ...) techniques and is highly extensible through packages.
Furthermore, R allows to produce well-designed publication-quality plots, including mathematical symbols and formulae, where needed.
R is available as Free Software under the terms of the Free Software Foundation's GNU General Public License in source code form. It compiles and runs on a wide variety of UNIX platforms and similar systems (including FreeBSD and Linux), as well as Windows and macOS.
Use R on bwUniCluster 3.0
The centrally maintained R module(s) should be preferred over alternatives like Conda or containerized setups when the computational performance is of major concern. The R installation(s) provided by these module(s) are built and optimized for the specific architecture and libraries of the cluster (e.g., OpenBLAS, MPI).
By contrast, if the focus is on portability of a data analytical workflow and maintaining many (external) dependencies, containerized setups or Conda may provide useful alternatives to the native R installation.
Furthermore, the R installation also provides the standalone library libRmath. This library allows you to access R routines from your own C or C++ programs (see section 9 of the 'R Installation and Administration' manual).
Loading the R software module
It is generally recommended to load a specific version of R, e.g.,
$ module load math/R/4.4.2-openblas
in data-analytical workflows instead of the default version which may change from time to time.
Installing R-Packages into your home folder
Since we cannot provide a software module for every R package, we recommend to install special R packages locally into your home folder. One option for doing this is from within an interactive R session:
> library() # List preinstalled packages
> install.packages('package_name', repos="http://cran.r-project.org") # Installing your R package and the dependencies
> library(package_name) # Loading the package into you R instance
The package is now installed permanently in your home folder and is available every time you start R.
Note:
By default R uses a version (and platform) specific path for personal libraries, such as "$HOME/R/x86_64-pc-linux-gnu-library/x.y" for R version x.y.z. This directory will be created automatically (after confirmation) when installing a personal package for the first time.
A version specific path, such as the default path, allows users to maintain multiple personal library stacks for different (major and minor) R versions and does also prevent users from mixing their stack with libraries built with different R versions.
The drawback is that, whenever switching to a new R release, the personal library stack must be rebuilt with that new R version into the corresponding (version specific) library path. This is considered good practice anyway in order to ensure a consistent personal library stack for any specific R version in use.
Pre-installed R-packages
- Rmpi
- iterators
- foreach
- doMPI
- doParallel
Installation instructions for selected R packages
The following guides provide detailed instructions for building selected optional R packages on bwUniCluster for R version 4.4.2. Please write a ticket if the instructions do not work for you or are outdated.