Description
This FREE workshop is sponsored by the University of Hawai?i Information Technology Service Cyberinfrastructure group and Hawaii Data Science Institute (HI-DSI)
This workshop focuses on Utilization of High performance Computing (HPC) clusters, such as Mana, for deep learning tasks as well as the benefits of understanding the different types of file systems available on HPC clusters and basic ways one would stage data from slower file systems to faster file systems.
Attendees will learn how to use Jupyter Lab on Mana via Open OnDemand, the benefits of using GPUs over just CPUs for deep learning applications, and how staging data in the correct file system can improve the performance of deep learning jobs.
Prerequisites
- Have an account on Mana
- MFA/DUO enabled on your UH Account: https://www.hawaii.edu/its/uhlogin/
- Be able to connect to the workshop in Zoom
- Have a modern web browser
- Basic understanding of Linux commands
Learning objectives
By the end of this workshop attendees will be able to:
- Have a basic understanding of how to access Open OnDemand and use Jupyter Lab on Mana
- Have a basic understanding of how to request GPUs and utilize them for deep learning tasks
- Have a basic understanding of the performance difference between CPU and GPU for deep learning
- Have a basic understanding of different types of file system found on HPC clusters
- Have a basic understanding of how correct file system choice affect deep learning performance
Tools Used for this workshop
- Open OnDemand
- Jupyter Lab