COUPLED 2025

Partitioning and load-balancing for a parallel solver of cell-by-cell cardiac electrophysiology

  • Trotter, James David (Simula Research Laboratory)
  • Cai, Xing (University of Oslo)

Please login to view abstract download link

A need for increasing physiological detail has led to mathematical models in cardiac physiology that are capable of resolving details at the level of individual cells. Increased physiological accuracy is accompanied by huge computational demands, and thus a need for high-performance computing. This talk presents a parallel numerical strategy based on domain decomposition for a cell-by-cell model, known as the EMI model. Due to separating extra- and intracellular domains, we propose suitable partitioning strategies to overcome load balancing and scalability challenges. Numerical experiments are presented to evaluate the solver’s performance and the impact of partitioning on its scalability.