My factor base is greater then yours (around two times).
I tried various factoring algorithms
- resieving
- trial division
- pollard rho
- mixture of resieving and trial division
- mixture of resieving and trial division with large primes
mixture of resieving and trial division (at the 500 th prime in the factor base i switch) give best results. The factoring needs half of the complete time (the other half is due to sieving).
So, not surprisingly sieving with large primes does not improve the speed.
For the example '732197471686198597184965476425281169401188191' given in
http://www.mersenneforum.org/showthread.php?t=7212.
I need 1.6 sec with my java application to find the factor.