High-Performance Computing (HPC) Series

September 21, 2022

You're Invited to the HPC Discussion Series hpc image

The High-Performance Computing (HPC) discussion series is open to all UCSC instructors, researchers, staff, and students.

Each series will include a special topic of interest designed to educate current and new users on HPC basics, how-to’s, best practices, time for open discussion, and Q&A.

DETAILS:

September 29, 2022 1:00 PM to 2:00 PM (PST)

Introduction to Hummingbird

The Nautilus Cluster is a highly distributed, but centrally managed community computational cluster. It is an on-demand, interactive machine with >500 GPUs (of all kinds) and >7000 CPU cores. Researchers on more than 30 campuses use Nautilus to run their software in containers, orchestrated by Kubernetes. This allows researchers to scale their computation from their laptop to dozens or even 100s of GPUs or CPUs at a time. Unlike, traditional High Performance Computing, Kubernetes-controlled clusters can be used interactively and scaled dynamically. This has major advantages for those interested in developing novel machine learning techniques or or deploying large-scale, interactive computational workflows. Additionally, the cluster has over 2PB of storage, attached at 10-100Gbps, so it can accommodate very large data sets.

Topics will include:

  1. Hummingbird System Overview
  2. Hummingbird Etiquette
  3. Software Module System
  4. Data Management
  5. Slurm Scheduler and Job Submission
  6. Checking Cluster and Job Status
  7. Getting Help

Date: Thursday September 29, 2022, 1:00pm - 2:00pm.

Zoom link: https://ucsc.zoom.us/j/93463299124?pwd=RFZvZzlYSjIzNnoxblFsRUV6aTZGZz09