MCell-K: A Highly Scalable Cell Microphysiology Simulator on Blue Gene L

Scott B. Baden, Gregory T. Balls, Terrence Sejnowski, Thomas M. Bartol

Abstract

MCell-K is a scalable variant of MCell, a highly successful Monte Carlo simulator of cellular microphysiology, which uses a random walk to model reaction/diffusion systems. We present new scalability results on hundreds of processors of Blue Gene and IBM Power4 systems, demonstrating exceptional scalability on a realistic simulation of neurotransmission. We discuss performance issues, including run time support and algorithmic changes needed to achieve high scalability.