GPU Cloud Bursting for Multi-Messenger Astrophysics with IceCube
Research perspectives from those at the cutting edge
On November 16, the Saturday before SC19 - the International Conference for High Performance Computing, Networking, Storage, and Analysis in Denver, the University of California (UC) San Diego performed a graphics processing unit (GPU) cloud burst during which they purchased the entire for-sale capacity of all GPUs available across Amazon Web Services (AWS) and other cloud providers worldwide.
Join this webinar to learn about the scientific motivation behind this project, hear about the preparation and execution, and learn about follow up bursts. In addition, you will discover the extent to which this is applicable to other sciences as well as commercial applications. At peak, UC San Diego operated a sustained workflow across 51,500 GPUs, eight generations of NVIDIA GPUs, and 28 cloud regions across three cloud providers. In a follow up burst, they integrated cloud and on-premises, simplifying the workflow.
When:Thursday, April 30, 2020 | 9:00 AM - 10:00 AM PT | 12:00 PM - 1:00 PM ET
Speaker:Frank Wuerthwein, Executive Director, Open Science Grid (OSG), and Professor of Physics, UC San Diego
Professor Wuerthwein is an experimental particle physicist who analyzes data from the Large Hadron Collider. He is the executive director of OSG responsible for providing an integrated data and compute platform that advances open science through distributed high throughput computing. Scientists used the platform in the detection of gravitational waves (Nobel Prize 2017), the discovery of the Higgs Boson (Nobel Prize 2013), and other scientific achievements.