Distribution of Random Streams in Stochastic Models in the Age of Multi-Core and Manycore Processors

Stochastic Modeling & Simulation require very good random sources which have now been available for more than a decade. However the parallelization of random streams remains delicate for many practitioners. Recent experience shows that a misuse of pseudo-random number parallelization techniques is not infrequent in various simulation domains. This talk will present the state of the art in the distribution of pseudo-random numbers and will also discuss the use of such techniques for hybrid computing with GP-GPUs.