Supercomputing Spotlights - Road to AI for Science
Society for Industrial and Applied Mathematics via YouTube
You’re only 3 weeks away from a new language
AI Engineer - Learn how to integrate AI into software applications
Overview
Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Discover the evolution of couplers in multi-physics simulations and their expanding role in AI-driven heterogeneous supercomputing through this 51-minute conference talk by Kengo Nakajima from the University of Tokyo and RIKEN R-CCS. Trace the historical development of coupler technologies while examining their connection to software sustainability challenges in advanced computing environments. Learn about the original function of couplers in linking complex models such as atmosphere-ocean and fluid-structure interactions, and understand how evolving workloads and hardware have transformed their applications. Explore insights from three major research projects: ppOpen-HPC (2011–2018) featuring MPI-based multi-physics couplers, h3-Open-BDEC (2019–2024) integrating Simulation, Data, and Learning (S+D+L) on Wisteria/BDEC-01 systems, and JHPC-quantum (2023–2028) advancing toward Quantum-HPC hybrid computing. Gain a comprehensive perspective on how couplers serve as bridges between simulation, artificial intelligence, and quantum computing technologies, positioning them as critical components in the future of scientific computing and AI for science applications.
Syllabus
Supercomputing Spotlights: Road to "AI for Science"
Taught by
Society for Industrial and Applied Mathematics