High Performance Computing and AI for Atmospheric and Oceanic Flows on Exascale Computers

S. Marras and Y. Tissaoui

Abstract:

The dawn of exascale computing presents unprecedented opportunities for advancing the fidelity and scope of atmosphere and ocean models. This minisymposium is meant to convene experts in high-performance computing and Earth system science to explore strategies for maximizing the potential of these powerful architectures. We will delve into cutting-edge techniques for parallelization, data management, and I/O optimization specifically tailored for complex climate and weather simulations. Presentations will highlight novel algorithmic developments, efficient utilization of heterogeneous architectures (CPUs, GPUs), and experiences porting and scaling models to exascale platforms. Contributions about both classical numerical approaches and hybrid-AI techniques are welcome. A key focus will be on addressing the unique challenges posed by the immense data volumes and computational demands of next-generation atmospheric and oceanic modeling systems. Join us to discuss the path forward in unlocking the full predictive capability of our planet through exascale innovation.

Simone Marras - New Jersey Institute of Technology, Department of Mechanical and Industrial Engineering

Simone Marras is Associate Professor of Fluid Dynamics at New Jersey Institute of Technology. He has been working in scientific computing and computational fluid dynamics for fifteen years and he enjoys writing open source software to solve PDEs.

Yassine Tissaoui - Barcelona Supercomputing Center, Computer Applications in Science and Engineering (CASE) Department

Yassine Tissaoui is a postdoctoral researcher at the University of Wisconsin-Madison. He has been working in computational fluid dynamics and scientific computing for eight years and atmospheric flow simulation for six years. He enjoys software development and always welcomes a technical challenge.