Iterative Methods for Earthquake Cycle Simulations with HPC Applications

Alexandre Chen
Date and time: 
Fri, Jun 10 2022 - 3:00pm
Alexandre Chen
University of Oregon
  • Brittany A. Erickson (Chair)
  • Jee Choi
  • Boyana Norris

High Performance Computing (HPC) is a powerful tool in scientific research for experimental simulation and data analysis. However, it has not been well studied and applied for earthquake cycle simulation. In this work, we explore computational methods and HPC techniques for earthquake cycle simulation with a focus on iterative methods and preconditioning techniques. The problem that we target is solving Poisson's equation numerically with boundary conditions derived from geophysics. Specifically, we use summation-by-parts (SBP) finite difference operators with simultaneous-approximation-term (SAT) method. The combined SBP-SAT scheme has high-order accuracy and provable stability for our problem. The purpose of this work is to examine existing research on problems similar to ours, so we can design algorithms for a large-scale 3D earthquake cycle simulation suitable for HPC platforms in the future. We begin with a mathematical introduction to our problem, and then iterative methods for solving this problem numerically. Finally, we explore the HPC aspects of these different iterative methods.