Thaler, FelixSolenthaler, BarbaraGross, MarkusJan Bender and Christian Duriez and Fabrice Jaillet and Gabriel Zachmann2014-12-162014-12-162014978-3-905674-71-2https://doi.org/10.2312/vriphys.20141230We present an architecture for parallel computation of incompressible IISPH simulations on distributed memory systems. We use orthogonal recursive bisection for domain decomposition and present a stable and fast converging load balancing controller. The neighbor search data structure is derived such that it optimally fits into the parallel pipeline. We further show how symmetry aspects of the simulation can be integrated into the architecture. Simultaneous communication and computation are used to minimize parallelization overhead. The seamless integration of these parallel concepts into IISPH results in near linear scaling for large-scale simulations.I.3.3 [Computer Graphics]Three Dimensional Graphics and RealismAnimationI.3.1 [Computer Graphics]Hardware ArchitectureParallel Processing