Quote:
Originally Posted by ATH
ewmayer already tested to k=2^{40} which is 2*2^{40}*97600541752017987211 which is 107.4 bits

Forgot to provide the run summary data:
Using first 1000000 odd primes; max gap = 154
max sieving prime = 15485867
searching in the interval k=[0, 1099516017600]
...
M(97600541752017987211) has 0 factors in range k = [0, 1099516017600], passes 015
Performed 37291183295 trial divides
Clocks = 24:57:19.683 hh:mm:ss.sss
This was on 1 CPU of a dualcore 2GHz Core 2 system (my vintage 2009 macbook).
The code also does spotchecks of candidate factors (make sure they don't have any factors within the range of the smallprime sieve; check if candidate q is a Fermat base2 PRP). For this size range of q's te code spotchecks every 2
^{21}st candidate q; here are those stats:
7105 base2 PRPs
10676 composites (with smallest factor > 15485867).