Umbrella HPC Service

This document is intented for faculty at institutions using the RCDS's Umbrella HPC service.

Getting Started

First log into our accounts management webpage to create your own account and accounts for students you supervise. Click the green 'New User' button and fill out the form on the next page. Next, click the blue 'add service' button and choose the account type under the 'New Service Request' heading. The various account types available are:

  1. Sandbox For users just starting out with minimal data retention requirements.
  2. Student For students in classes that need HPC resources.
  3. Thesis For students working on thesis research projects.
  4. Research For students or faculty doing longer term research projects.

A temporary password will be emailed to the address you enter. If you need to create many users at once, email us and we can do this in a batch process.

What Services are available?

RCDS offers High Performance Compute resources and common data analysis software on a linux platform. A description of the hardware resources available is here. You can access web-based IDE's such as JupyterHub and RStudio on our standalone servers. You can also SSH in to the standalone servers for command line interface (CLI) access. The web-based IDE's are at:

JupyterHub : https://standalone_server_name.ibest.uidaho.edu:8000
Rstudio    : https://standalone_server_name.ibest.uidaho.edu

The JupyterHub web application also allows you to interact with the CLI through a terminal session hosted in a web page (click New -> Terminal).

We have a large variety of bioinformatics and general research software installed, when logged in the 'module avail' command will show you what is available.

The cluster is available only through the CLI. We have a workshop that is a good introduction to cluster computing.

Getting data to the server

With the exception of onyx.ibest.uidaho.edu (sandbox accounts), all our servers share a distributed file storage system which means that it doesn't matter which server you upload data to - it will end up in the same place.

To download data directly from the internet, use wget from a terminal session, for example:

wget -nc http://www.hpc.uidaho.edu/example-data/Myco.tb.fasta

If your data is on your local computer, can upload it through the JupyterHub IDE, or you can scp the data to the server:

scp /path/to/local/data user@server.domain.com:/path/to/destination

Note: SCP does not deal well with spaces in paths or filenames.

Other graphical SFTP client options:

Questions?

Contact us