• Home page: The R Project for Statistical Computing
• Licenza: GPL

Several versions of R are available:

Compilation details:

For example the three versions of R 4.1.2 have been compiled with:

• GNU 5.4.0 tool chain and OpenBLAS 0.2.20
• GNU 7.3.0 tool chain and OpenBLAS 0.2.20
• GNU 8.3.0 tool chain and OpenBLAS 0.3.7

Environment modules:

The R module defines the environment variable

R_LIBS_USER

with the default value

$HOME/local/libraries/R/<R_MAJOR_VERSION>.<R_MINOR_VERSION>

So if you intend to install additional R packages, you have to create the directory where the libraries will be installed:

mkdir -p "$R_LIBS_USER"

As you can see the value of R_LIBS_USER does not include the patch level of the software.

This means, for example, that the packages installed in

$HOME/local/libraries/R/4.1

will be suitable for all versions of R 4.1.x series.

In this example we will see how to install two R packages.

Script install-packages.R:

install.packages("parallel", repos="http://cran.mirror.garr.it/mirrors/CRAN")
install.packages("MASS"    , repos="http://cran.mirror.garr.it/mirrors/CRAN")

Script install-packages.sh:

install-packages.sh

#!/bin/bash
 
module load gnu8/8.3.0
module load R/4.1.2
 
mkdir -p "$R_LIBS_USER"
 
R CMD BATCH install-packages.R
 
rm -f install-packages.Rout

Executing install-packages.sh:

bash install-packages.sh

In this example we will see how to submit a serial job.

R script hello.R:

hello.R

#!/usr/bin/env Rscript
 
sayHello <- function(){
   print('hello')
}
 
sayHello()

Script slurm-R-vrt.sh:

slurm-R-vrt.sh

#!/bin/bash
#SBATCH --job-name=hello
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=1
#SBATCH --mem=512M
#SBATCH --partition=vrt
#SBATCH --time=0-00:05:00
##SBATCH --account=<account>
 
module load gnu8/8.3.0
module load R/4.1.2
 
Rscript --no-save hello.R

Submitting slurm-R-vrt.sh:

sbatch slurm-R-vrt.sh

In this example we will see how to submit a parallel job on a single node.

R script loop.R:

loop.R

#!/usr/bin/env Rscript
 
library(parallel)
library(MASS)
 
args = commandArgs(trailingOnly=TRUE)
 
if (length(args)==0) {
  args[1] = 1
}
numCores <- args[1]
 
x <- iris[which(iris[,5] != "setosa"), c(1,5)]
trials <- seq(1, 10000)
boot_fx <- function(trial) {
  ind <- sample(100, 100, replace=TRUE)
  result1 <- glm(x[ind,2]~x[ind,1], family=binomial(logit))
  r <- coefficients(result1)
  res <- rbind(data.frame(), r)
}
system.time({
  results <- mclapply(trials, boot_fx, mc.cores = numCores)
})

Script slurm-R-cpu.sh:

slurm-R-cpu.sh

#!/bin/bash
#SBATCH --job-name=loop
#SBATCH --output=%x.o%j
#SBATCH --error=%x.e%j
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=4
#SBATCH --mem=512M
#SBATCH --partition=cpu
#SBATCH --time=0-00:05:00
##SBATCH --account=<account>
 
module load gnu8/8.3.0
module load R/4.1.2
 
Rscript --no-save loop.R ${SLURM_NTASKS_PER_NODE:-1}

Submitting slurm-R-cpu.sh:

sbatch slurm-R-cpu.sh