We describe an algorithm for dynamic load balancing of geometricallyparallelized synchronous Monte Carlo simulations of physical models. Thisalgorithm is designed for a (heterogeneous) multiprocessor system of the MIMDtype with distributed memory. The algorithm is based on a dynamic partitioningof the domain of the algorithm, taking into account the actual processorresources of the various processors of the multiprocessor system.
展开▼
