Michael Mascagni
Serial and Parallel Random Number Generation: Theory and Practice

Providing high quality pseudorandom numbers for parallel computers supplies many deep and fascinating mathematical problems as well as unique software engineering challenges. We will give a brief introduction to both pseudo- and quasirandom number generation. We will then present some techniques for parallel pseudorandom number geneation based on our own research. One of the more practical issues in parallel pseudorandom number generation is finding methods that provide portability and reproducibility across architectures. The techniques described provide good quality parallel random numbers that have many desirable properties, including portability and reproducibility across architectures. These generators form the basis for a DARPA sponsored project that has developed a scalable library for pseudorandom number generation (SPRNG) in collaboration with researchers at the University of Illinois, Urbana-Champaign. We then describe the SPRNG library in some detail, and give a brief tutorial in the use of SPRNG for both serial and parallel Monte Carlo computations.