Fast and reliable random number generators

217. R. P. Brent, Fast and Reliable Random Number Generators for Scientific Computing (invited talk), Proceedings of the PARA'04 Workshop on the State-of-the-Art in Scientific Computing, Lyngby, Denmark, 20--23 June 2004, Lecture Notes in Computer Science, Vol. 3732 (2006), 1-10. Extended abstract in the preliminary Proceedings (CDROM), Volume 1, 3-9.

Extended abstract: dvi (14K), pdf (153K), ps (58K).

Preprint: dvi (19K), pdf (177K), ps (69K).

Paper: Available from Lecture Notes in Computer Science.

Overhead Transparencies: dvi (22K), pdf (240K), ps (152K).


Fast and reliable pseudo-random number generators are required for simulation and other applications in Scientific Computing. We outline the requirements for good uniform random number generators, and describe a class of generators having very fast vector/parallel implementations with excellent statistical properties. We also discuss the problem of initialising random number generators, and consider how to combine two or more generators to give a better (though usually slower) generator.


See [211] for a related paper and [185] for a survey of random number generators.

Go to next publication

Return to Richard Brent's index page