Using R on Quest

Availability

R is available for use on Quest via a web browser, or for use with scheduled jobs.

Scheduled Jobs: R Versions and Basic Use

You can see which versions of R are available on Quest, and which version is the default, with the command

module avail R

You can make a particular version of R available to use by typing the full module name with the version included as listed in the output from module avail R.  Example:

module load R/3.3.1

You need to load an R module even if you want to use the default version of R. 

After you've loaded the R module, you can use the command R to start the command line version of R or rstudio to launch the graphical IDE. If you want to use RStudio, you need X-windows functionality enabled for your connection. See Logging in to Quest for instructions on connecting to Quest with support for the X Window System.

Batch Jobs

In submission scripts for batch jobs, you must include the command to load the version of R that you want to run, then you can call R using:

Rscript myprog.R

which causes the output and errors to be written to the files for stdout and stderr, respectively, which are determined by your job submission parameters (see Submitting a Job on Quest).

See Rscript documentation for additional options; you may want to use the --verbose option. Note that unlike R, Rscript does not load the methods package by default.  If your script works interactively or on your system, but gives you errors as a batch job, try adding library(methods) to the top of your script.

You should install any packages that your script needs prior to submitting your batch job. It is recommended to start R interactively from a login node to install packages so that you can monitor the installation and check for any errors.

Interactive Use

To use R (likely via RStudio) interactively, please submit an interactive job to obtain the memory and cpu resources you need. R and RStudio may be run directly on the login nodes when installing packages or testing small scripts, but computationally intensive work should not be done on the login nodes.

Usage Notes

While module load R will load the current default version of R, to ensure the future compatibility of your scripts with the Quest system, we highly recommend that you always load a specific version of R rather than relying on the default version.

If using RStudio, know that RStudio automatically creates and saves a .Rhistory file in your default working directory when you quit. If you do not change the working directory during your R session, then the default working directory is the directory from which you launched RStudio. RStudio also automatically loads any available .Rhistory file from the default working directory when it starts. .Rhistory files will be accessible to, and possibly shared by, anyone with access to the directory. You can prevent RStudio from automatically saving the history file by going to the Tools > Global Options menu in RStudio. The command line version of R only creates and saves an .Rhistory file when creating a .RData file.

Installing and Managing R Packages on Quest

R packages are installed, whether system-wide or locally in your own directory, for specific versions of R. If you switch versions of R, you may need to reinstall packages you use. However, packages only need to be installed once on Quest for each version of R.  Once installed, a package is available across all log in, compute, and analytics nodes on Quest for that version of R.

System-wide Packages

A limited number of R packages have been installed centrally for each version of R. These packages can be used without downloading and installing them first. You can get a list of the currently installed packages with the command:

installed.packages()

Package Installation

Users can install additional R packages from CRAN by launching RStudio and using the Packages tab in the bottom-right window, or by running the R command line console and using the install.packages command:

install.packages(c("packagename1", "packagename2"))

with a list of the packages you need. Packages should generally be installed under your home directory. They will be available for you to use across Quest nodes.

First-time Installation

The first time you install an R package on Quest, you may see an error like:

 Warning in install.packages("glmnet", repos = "http://cran.r-project.org") :
  'lib = "/hpc/software/R/3.3.1/lib64/R/library"' is not writable
  Would you like to use a personal library instead? (y/n)

Answer "y" to use a personal library.  Then it will ask you:

 Would you like to create a personal library
  ~/R/x86_64-pc-linux-gnu-library/3.3
  to install packages into? (y/n)

Answer "y" again. The installation should then proceed successfully. This will install the packages in your home directory.

Setting the Repository

If you get an error message indicating that a particular package isn't available for a certain version of R, or you get another error related to the package repository, you may need to explicitly set the mirror from which you want R to download the package. List of CRAN mirrors. Choose an http (instead of https) mirror. You can set the mirror before trying to install packages with the command:

options(repos="http://cran.wustl.edu")

or as part of the install.packages call with the repos argument:

install.packages(c("packagename1", "packagename2"), repos="http://cran.wustl.edu")

Compiled Packages

For Use on Login and Compute Nodes

Some packages compile underlying code as a part of the installation. To install such packages, you may need to load a newer version of the gcc compiler before starting R and installing the package. Loading a different compiler should NOT be done for packages you intend to use on the Analytics nodes, as the updated gcc modules are not available on those nodes.

For example, if while installing a package, you see an error like:

configure: WARNING: Only g++ version 4.6 or greater can be used with...

try loading the module for a newer version of gcc; For example, after logging into Quest:

module load gcc/5.1.0
module load R/3.3.3
R

then in the R console, issue the package installation commands again.

For Use on Analytics Nodes

R packages intended for use on the Analytics nodes should be installed from within an R session running in RStudio in the web browser on the Analytics nodes. If you encounter compilation or installation errors for a package, please contact quest-help@northwestern.edu with the details for further assistance.

R packages installed via the Quest login or compute nodes with the module R/3.5.1 will also be available on the Analytics nodes.

Note

Some compiled packages cannot be successfully installed by individual users. Please contact quest-help@northwestern.edu if you run into trouble with a particular package.

More Options

See the R help page for install.packages for more options, such as installing local (user-created or downloaded) packages or changing the directory where packages are installed. The latter can be used to install packages in your project space instead of your home directory, which might be useful if others on the allocation also need to access the packages. To install packages from GitHub, use the install_github function in the devtools package or use the githubinstall package.

Task Views

CRAN Task Views are collections of packages used frequently in particular fields. They allow you to install all of the packages associated with the task view at once. See the R documentation for more details and the commands to use.

Additional Help

Please contact quest-help@northwestern.edu for additional help with R package installation.

See Also:




Keywords:quest, R, statistics, statistical, computing, graphics, stats, install, package   Doc ID:71270
Owner:Research Computing .Group:Northwestern
Created:2017-03-02 16:27 CDTUpdated:2019-04-15 13:21 CDT
Sites:Northwestern
Feedback:  1   0