Parallel Simulation of Multicomponent Systems

Mike Heath, University of Illinois

Complex engineering systems and other physical phenomena often involve multiple components whose interactions are crucial to determining the overall behavior of the system. Simulation of such multicomponent systems offers numerous challenges in mathematical modeling, geometric representation, algorithm design, and software engineering. Moreover, the enormous computational demands of such simulations usually require implementation on large-scale parallel computers, which introduces further issues such as partitioning, load balancing, scalability, etc. In this talk we will address many of these issues, as exemplified by whole-system simulation of solid propellant rocket motors on massively parallel computers.