By Topic

Computing twin primes and Brun's constant: a distributed approach

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

3 Author(s)
P. H. Fry ; Dept. of Comput. Sci., Rensselaer Polytech. Inst., Troy, NY, USA ; J. Nesheiwat ; B. K. Szymanski

The paper describes an implementation of a large heterogeneous distributed parallel computation that counts the distribution of twin primes and calculates Brun's constant and maximal distances between pairs of twin primes. Two primes are twins if they differ by two. It is not known if there are infinitely many twin primes but it was proven that the sum of their inverses converges to the value defined as Brun's constant. Prior to this work, the number of twins and their contribution to Brun's constant was known for all twins up to 1014. The authors have advanced this calculation to 2·1015 and are planning to continue to 1016. The computation is distributed using the farmer-workers paradigm. The farmer divides the numerical range into intervals of 10 billion. Workers are assigned intervals to analyze and return results. If a worker fails, e.g., it does not return results within a given time period, the farmer reassigns its intervals to other workers. The farmer also frequently saves its state. If the farmer dies, a new one can be started up with minimal loss of work. The paper describes the algorithm for locating twin primes, a modified Sieve of Eratosthenes. It also discusses the method for distributing the work to clients, time and space saving optimizations, and multi-platform support. They conclude by presenting some preliminary results up to 1.8·1015 future directions for the system

Published in:

High Performance Distributed Computing, 1998. Proceedings. The Seventh International Symposium on

Date of Conference:

28-31 Jul 1998