The Cunningham Project seeks to factor so-called “Cunningham numbers”, these are numbers of the form bn ± 1, for b = 2, 3, 5, 6, 7, 10, 11, 12 and n small.
Unfortunately, it seems that the tables from the Cunningham Project are not meant to be parsed by a computer program (actually, while creating this page I found out about Paul Leyland's tables, which are much better). That's why I decided to create this page. It contains the Cunningham Tables and some extensions in a machine-readable format, as well as lists of the primes and composites occuring in the tables. These tables are also incorporated in my fullfactor package for PARI/GP.
These tables contain the factorizations of bn ± 1, including the non-primitive (algebraic) part. In the case of Aurifeuillian (L·M) factorizations, the L and M parts are not separated. If for certain b and n, both bn - 1 and bn + 1 are in the table, then the implied factorization of b2n - 1 is also given. All this also explains why some numbers in my tables have multiple unfactored composites.
My tables were last updated on 2014-03-30, amalgamating all data from the following sources:
Occasionaly, I will also add factors by hand, if I happen to learn about a new factorization.
Instead of trying to parse any of this, I just extract all numbers using some sed hacking. Then I use PARI/GP to figure out which Cunningham numbers are multiples of the factors, from there it is easy to make the tables. This whole process takes about an hour. There are sanity checks along the way to make sure the data is correct (the worst that could happen is that my tables list a composite for which the factorization is already known).
Everything is completely automatic, from the downloading of the sources, to updating this webpage. My tables should be remade within a day whenever one of the sources changes.
These give the factorizations of bn ± 1. Lines are of the form
This example gives the partial factorization of 281 + 1, with angle brackets around composites. All other factors are BPSW pseudoprimes (using PARI/GP's ispseudoprime() function).
Download main.gz (1 618 678 bytes gzip'ed text file).
This is a list of all prime factors > 500 000 occuring in the Main Tables. Currently, there are 22926 primes in this table. Lines are of the form
This means that 52370346649565455417091 is a prime factor of 3115 - 1. Obviously, this prime also divides 3230 - 1, but in the file every factor is given only once per base. However, the prime 1505447 (and others) occurs twice in this file, because it divides 2173 - 1, as well as 6229 - 1.
Download primes.gz (572 939 bytes gzip'ed text file).
This lists the 407 composites in the Main Tables, which have not been factored yet. The format is the same as the primes table above.
Download composites.gz (55 600 bytes gzip'ed text file).
- Sam Wagstaff's page: The ‘official’ resource about the Cunningham Project. Without this, I couldn't have made my tables.
- Paul Leyland's page: Also has some nice tables with factorizations of the Cunningham numbers.
- NFS@Home: A distributed computing project to factor Cunningham numbers using the Number Field Sieve.
- ECMNET: Another distributed computing project, using the Elliptic Curve Method to find ‘small’ factors.
- mersenneforum.org: A forum about the GIMPS project, as well as other factoring projects mentioned above.