For the most time-consuming number
crunching, we have to
effectiveness of two programs, namely Paul Jobling's "NewPGen" and Jean
Penne and George Woltman's "LLR".
NewPGen is the sieving
program we use to check all 321 candidates
small divisors. For example, on testing for divison by 5, a quarter of
all candidates will be removed as not prime. However it becomes
increasingly time-consuming to find new divisors of any of the
candidates and we switch over to the program LLR which
the current level of testing, determine the primailty of any remaining
candidate in a few hours.
As of 16th March 2006, after three years of
sieving on many computers, we have reached the level where it is now
quicker to just run LLR on all remaining prime candidates.
Early 2007 update: due to improved software we are able to sieve a little further with srsieve.
to 321 index page