Actions

Difference between revisions of "R"

From Montana Tech High Performance Computing

(RStudio Server)
 
(9 intermediate revisions by the same user not shown)
Line 1: Line 1:
The R package for statistical computing and graphics is installed on the system. Version 3.3.3 installed from source code is available and installed in <code>/opt/R/R-3.3.3</code>. To access it:
+
The R package for statistical computing and graphics is installed on the system. Version 3.6.1 installed with the gnu8 toolchain. To access it:
  
: <code>module load R/3.3.3</code>
+
: <code>module load R</code>
  
An example job script, might look something like:
+
===Install R Packages===
 +
Generally, if you need to use R packages that are not available, you can always install it locally in your account.
  
: <code style=display:block>#!/bin/sh<br>#PBS -l nodes=1:ppn=32<br>#PBS -N PingJob<br>#PBS -m e<br>#PBS -M username@mtech.edu<br>#PBS -l walltime=00:01:00<br><br>cd $PBS_O_WORKDIR<br>module load R/3.3.3<br>R < parLapply_test.R > parLapply_test.output --no-save</code>
+
First start an R session:
 +
 
 +
:<code style=display:block>module load R<br>R</code>
 +
 
 +
Then you can use the <code>install.packages</code> command inside R to install a package. For example,
 +
:<code style=display:block>> install.packages("ggplot2")</code>
 +
Because you don't have a write permission in the /opt directory, where R is installed, R will throw the following warings:
 +
:<code style=display:block>Warning in install.packages("ggplot2") :<br>   'lib = "/opt/ohpc/pub/libs/gnu8/R/3.6.1/lib64/R/library"' is not writable<br>Would you like to use a personal library instead? (yes/No/cancel)</code>
 +
Answer <code>yes</code>, R will then ask you to create a library folder in your own directory:
 +
:<code style=display:block>Would you like to create a personal library<br>‘~/R/x86_64-pc-linux-gnu-library/3.6’<br>to install packages into? (yes/No/cancel)</code>
 +
Answer <code>yes</code> again, R will start the installation process.
 +
 
 +
In addition, you can also install R packages in a different folder other than the default one. For example, to install a package at <code>~/libs/R_libs</code>:
 +
 
 +
First create the directory:
 +
:<code style=display:block>mkdir ~/libs/R_libs</code>
 +
Then inside R:
 +
:<code style=display:block>> install.packages("ggplot2", lib="~/libs/R_libs")</code>
 +
===Submitting R jobs on HPC===
 +
 
 +
===RStudio Server===
 +
RStudio Server is installed on Copper, which provides an interactive RStudio experience via your web browser.
 +
If you have an account on Copper, you can access it from: http://copper.mtech.edu:8787
 +
 
 +
You can only access it on campus or via VPN.

Latest revision as of 10:33, 28 May 2020

The R package for statistical computing and graphics is installed on the system. Version 3.6.1 installed with the gnu8 toolchain. To access it:

module load R

Install R Packages

Generally, if you need to use R packages that are not available, you can always install it locally in your account.

First start an R session:

module load R
R

Then you can use the install.packages command inside R to install a package. For example,

> install.packages("ggplot2")

Because you don't have a write permission in the /opt directory, where R is installed, R will throw the following warings:

Warning in install.packages("ggplot2") :
'lib = "/opt/ohpc/pub/libs/gnu8/R/3.6.1/lib64/R/library"' is not writable
Would you like to use a personal library instead? (yes/No/cancel)

Answer yes, R will then ask you to create a library folder in your own directory:

Would you like to create a personal library
‘~/R/x86_64-pc-linux-gnu-library/3.6’
to install packages into? (yes/No/cancel)

Answer yes again, R will start the installation process.

In addition, you can also install R packages in a different folder other than the default one. For example, to install a package at ~/libs/R_libs:

First create the directory:

mkdir ~/libs/R_libs

Then inside R:

> install.packages("ggplot2", lib="~/libs/R_libs")

Submitting R jobs on HPC

RStudio Server

RStudio Server is installed on Copper, which provides an interactive RStudio experience via your web browser. If you have an account on Copper, you can access it from: http://copper.mtech.edu:8787

You can only access it on campus or via VPN.