This paper describes computational techniques for concurrent Direct Simulat
ion Monte Carlo (DSMC) of neutral flow inside three-dimensional plasma reac
tors. These techniques are designed to reduce the overall time to obtain re
alistic three-dimensional results. A simulation problem is first defined by
automatically generating a tetrahedral grid, This grid is statically parti
tioned in order to achieve scalable memory utilization. Parallel algorithms
allow high-performance multicomputer architectures to be leveraged to redu
ce simulation time, Adaptive gridding techniques are employed to maintain t
he desired accuracy in the presence of dynamic flow variations, Load balanc
ing algorithms, based on heat diffusion, are used to maximize efficiency. F
inally, automatic granularity control is used to ensure appropriate granula
rity for load balancing and for maximum processor utilization. Simulation r
esults are presented for a simulation of the GEC reference cell in a flow c
onfiguration of industrial relevance. Performance results are presented for
large-scale multicomputers, symmetric multiprocessors and networks of PCs,
Copyright (C) 2000 John Wiley & Sons, Ltd.