Thread: Octoproths
View Single Post
Old 2005-04-07, 12:04   #6
axn's Avatar
Jun 2003

490210 Posts

I used NewPGen BiTwin option which sieves for the 4 forms: k.2^n+/-1 and k.2^(n+1)+/-1

This reduces the billions of k's to about 10000. It is then run thru LLR, first for k.2^n+1, whose output is again LLR-ed for k.2^n-1, whose output is *again* LLR-ed for k.2^(n+1)+1, and finally (you guessed it!) LLR-ed for k.2^(n+1)-1

The resulting k's are then normalized (I sieve for even k's also) to odd k's. Then after a bit of script magic and Dario Alpern's batch factorization, I reduce the list to 2^n+k primes, then to 2^n-k primes, and finally the last two forms to come up with the solutions.

Incidentally, just completed n=66 and no solutions

Oh well. On to 67 !!

Last fiddled with by axn on 2005-04-07 at 12:16
axn is offline