Get Started with HPC
Are you new to high-performance computing (HPC), new to our services specifically, or just need a refresher on where to start? The "Get Started" section of our documentation site provides a general flow for getting access to our cyberinfrastructure, through submitting a job and viewing the end results.
At the bottom of each page, you'll be able to click right into the 'next step'.
The Steps
1. Request Access
By default, faculty and students are not given access to use the BOSE supercomputing cluster without first requesting access. This is done so we can ensure that everyone is set up properly with the right classification (research vs class use) and group permissions.
Use of the older BGSC supercomputing cluster is open to all UWEC accounts automatically, though we still request you reach out to us first.
2. Accessing Clusters
Once you have been given access to log into our BOSE or BGSC cluster, you'll be able to log in via SSH (1) or using your web browser through Open OnDemand.
- SSH access is only provided to those with access to a UW-Eau Claire account. If you are using credentials from another UW campus that was issued an HPC account, you will only be able to use our web-based applications such as JupyterHub.
3. File Transfers
Most scripts require input files, such as datasets, to be able to do any calculations or analyses. We have a variety of different methods to transfer files onto our computing clusters, such as through your web browser, SFTP, or via Globus for large files.
4. Submit Job
To handle submitting scripts to our computing cluster, we use a job scheduling tool called Slurm. Slurm allows you to specify the computational resources you require to run your script, and puts it in queue to automatically find space on one of our 65+ computational nodes.
5. View Job Results
Once your job finishes running, you'll be able to access any output and error logs to view your program's results. You'll also have the opportunity to see how well you utlized your resources requested (CPUs, memory, GPU), which we recommend reviewing so you can adjust for your next submission.
Acknowledgements
Have you successfully used our computing clusters for your research project and are now in the process of writing a paper, poster, or giving a presentation? We request that you acknowledge using the Blugold Center for HPC's resources, which helps advertise our services to other researchers.
Already had a paper published? Then let us know so we can advertise it our our website!
Stuck or would like some advice?
Are you just getting started and not sure what to do next? The HPC Team is available to meet directly with you and your group to walk you through each part of the process, tailoring it specifically to the research or class project you have in mind.