AI-coupled HPC Workflows: A New Paradigm for Extreme Scale Scientific Applications

ABSTRACT

Extreme-scale computing faces several significant inflection points. On the one hand, modern computing trends present challenges to scaling large, monolithic tasks. On the other hand, scientific applications increasingly demand seamless orchestration of diverse, heterogeneous tasks. In particular, workflows integrating AI and traditional HPC are gaining prominence and positioned to drive the next wave of “algorithmic Moore’s Law.” To meet these demands,extreme-scale computing platforms must transform, elevating workflows to first-class application entities. We will discuss coupled and integrated workflow applications across multiple scientific domains and the evolving software systems that enable them.

BIOGRAPHY

Shantenu is a Professor of Computer Engineering at Rutgers University-New Brunswick and the Head of Computational Sciences at DOE's Princeton Plasma Lab (while holding a concurrent appointment as a Research Scholar in Computer Science at Princeton University). His research interests lie in AI for Science, situated at the intersection of high-performance distributed computing and computational and data science. He has received the NSF CAREER Award (2013),the ACM Special Gordon Bell Award (2020), and numerous other awards in Supercomputing, and was the winner of the IEEE SCALE2018.