We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Numerical Analysis

Title: Millions of Perrin pseudoprimes including a few giants

Abstract: The calculation of many and large Perrin pseudoprimes is a challenge. This is mainly due to their rarity. Perrin pseudoprimes are one of the rarest known pseudoprimes. In order to calculate many such large numbers, one needs not only a fast algorithm but also an idea how most of them are structured to minimize the amount of numbers one have to test. We present a quick algorithm for testing Perrin pseudoprimes and develop some ideas on how Perrin pseudoprimes might be structured. This leads to some conjectures that still need to be proved. We think that we have found well over 90% of all 20-digit Perrin pseudoprimes. Overall, we have been able to calculate over 9 million Perrin pseudoprimes with our method, including some very large ones. The largest number found has 3101 digits. This seems to be a breakthrough, compared to the previously known just over 100,000 Perrin pseudoprimes, of which the largest have 20 digits. In addition, we propose two new sequences that do not provide any pseudoprimes up to $10^9$ at all.
Subjects: Numerical Analysis (math.NA); Number Theory (math.NT)
MSC classes: 11B37, 11B39, 11B50
Cite as: arXiv:2002.03756 [math.NA]
  (or arXiv:2002.03756v1 [math.NA] for this version)

Submission history

From: Holger Stephan [view email]
[v1] Mon, 10 Feb 2020 14:02:46 GMT (34kb,D)

Link back to: arXiv, form interface, contact.