Software on Lawrence
The Lawrence HPC uses modules for the use of pre-installed software applications on the HPC. Modules allow various versions of software to be available. This document provides a brief overview of module commands as they pertain to USD's Lawrence HPC.
Once you are logged into Lawrence, the module commands will be available to you. Following is a brief overview of the basic module commands:
Modules are for software that have been installed on Lawrence as a whole. If you have installed software in your own account (e.g. through Anaconda/Bioconda), it will be ready without being loaded.
The following example uses module commands to load R version 3.4.1 and then open R for command line use on Lawrence:
Anaconda
Anaconda is a channel software manager of conda which can be used for installing various applications. Anaconda is recommended for those wishing to install software that is not readily available as a module. More information on using Anaconda can be found on the Anaconda documentation page here. Packages currently available on Anaconda can be found here. Anaconda provides python by default. R may be installed using the command "conda install R".
To install Anaconda in your home directory on Lawrence, connect to a node, then run the install-anaconda script as follows:
Then run the installer. Make sure to select "yes" when asked about prepending the Anaconda3 install location to PATH.
Click "enter" to continue through the license. (You may need to click "enter" multiple times.)
Accept the terms:
Then click "enter" to accept the installation location when asked.
Let the installer initialize Miniconda3.
Do you wish the installer to initialize Miniconda3 by running conda init? [yes|no] [no] >>>yes
You may need to re-open your terminal for changes to take effect, before continuing
Check to make sure that conda is installed:
Then update your Miniconda to the latest version of Anaconda3:
Note: conda update conda
may also be used to update Anaconda/Bioconda later.
Packages
Packages that do not automatically come with Anaconda may be installed using the "conda install package-name" command. Package options are available here. R language package options are available here (see the "R" section for installation details for R packages).
R
After Anaconda is installed, to install R, use "conda install R":
MRO (Microsoft R Open) is also available, and provides threading.
Anaconda may also be used to install R packages. One popular package at USD is R-essentials. To install this, use the conda install command "conda install -c r r-essentials":
Bioconda
Bioconda is a version of Anaconda (a channel software manager of conda) that is designed especially for use in the biological sciences. It can be used for installing various bioinformatics applications. Bioconda is recommended for those wishing to install software that is not readily available as a module, especially in the biological sciences. More information on using Bioconda can be found on the Bioconda documentation page here. Packages currently available on Bioconda can be found here.
To install Bioconda in your home directory on Lawrence, connect to a node, then run the install-bioconda script as follows:
There will be a .sh file in your home directory. Run this using bash.
Press enter to continue, and to scroll through the license, then type "yes" to accept the license.
Press enter to confirm the installation location.
Type "yes" to allow the installer to run conda init
Wait for the installation to finish. Your command prompt will reappear to show that the computer is ready.
Log out of the node and Lawrence, then back into Lawrence for the changes to take effect.
Bioconda will now be available through your home directory to install software. It will appear as "Anaconda"
After Bioconda is installed, to install R, use "conda install R" (see instructions above: Anaconda - R)
Math Libraries
Intel Math Kernel Library (MKL)
The Intel MKL library is available on Lawrence and is the recommended math library for most applications. This library is configurable to various compilers and languages. Functions provided by Intel MKL include BLAS, LAPACK, and FFTW. For more information on Intel MKL, please visit the developer documentation webpage.
Intel also provides a tool to generate appropriate command line options for your particular compiler, integer size, and threading model which can be found here.
Example command line options for C and GCC compilers can be found below. These commands would be added after running "module load intel":
Intel C compiler:
GCC C compiler:
Last updated