首页

the past two decades, large-scale numerical simulations on parallel CPU clusters have been crucial for our theoretical understanding of star formation and stellar feedback. As modern supercomputers increasingly pivot toward GPU-dominated architectures, it has become essential to redesign our simulation tools to this new architecture in order to scale up these simulations. In this talk, I will give an overview of the state of the art in numerical methods for star formation simulations. I will then introduce my work co-developing QUOKKA, a next-generation, GPU-native, adaptive mesh refinement radiation-MHD code designed to scale from protostellar environments to galactic disks. I will present initial production results and outline a long-term roadmap that I am leading with QUOKKA. This roadmap aims to confront JWST-era benchmarks, spanning star formation across cosmic time, galactic winds, nuclear star cluster dynamics, and potential seeding channels for supermassive black holes.

BIO

ChongChong He is a Postdoctoral Fellow at the Research School of Astronomy & Astrophysics at The Australian National University. Dr. He earned his Ph.D. in Astronomy from the University of Maryland, College Park in 2023. Prior to that, he graduated with highest honors in Physics (B.S.) from Jilin University in 2016. His research has been focused on computational methods and numerical studies of star formation and feedback processes. He is currently developing GPU-optimized numerical methods for star formation and galaxy evolution simulations. His work has been recognized with well-regarded grants, including the Future Investigators in NASA Earth and Space Science and Technology award, as well as millions of GPU-hours of allocations on leading supercomputing facilities in Australia and the United States. More of his work can be found at his academic website:https://chongchonghe.github.io

Host: Hui Li