![]()
Note: This article is available in a single file for downloading and printing.
We address the following user questions in this article:
We do not intend this article to be technical, but aim at answering the above questions by referring to technical references. We especially want to call attention to ACM article [8] on the subject of random number generators.
The present article does not discuss what random numbers are or how they are used, and we do not single out any vendor's software as superior. We restrict our attention to single-precision floating-point generators, although integer generators also enter the picture in this context. We do not discuss third-party software offerings of random number generators other than the ones in [8].
A sequence of pseudo-random numbers is a sequence of numbers that satisfies certain properties of "randomness." Such sequences are repeatable, depending upon an initial seed. For complete details on the theory and testing of randomness, see [6].
For the purpose of this note, we use the term "random number" to mean "uniform random number" or "pseudo-random number". We also use the term "generator" to mean "random number generator."
A list of references accompanies this article.
Thanks go to Linda Mearns, a senior scientist in NCAR's Climate and Global Dynamics Division, for the inspiration for this article.