On pseudorandom number generators
Statistical sampling and simulations produced by algorithms require fast random number generators; however, true random number generators are often too slow for the purpose, so pseudorandom number generators are usually more suitable. But choosing and using a pseudorandom number generator is no simple task; most pseudorandom number generators fail statistical tests. Default pseudorandom number generators offered by programming languages usually do not offer sufficient statistical properties. Testing random number generators so as to choose one for a project is essential to know its limitations and decide whether the choice fits the project’s objectives. However, this study presents a reproducible experiment that demonstrates that, despite all the contributions it made when it was first published, the popular NIST SP 800-22 statistical test suite as implemented in the software package is inadequate for testing generators.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).