Trial Divison Improvement
I also wanted to mention the trial division of numbers (of special forms) in PFGW for (a^nb^n)/(ab) (with prime n) for integer roots a, b should be reduced to only trial dividing numbers of the form 2kn+1 (or if it is faster, sieving values of 2kn+1 and then trial dividing). Although little interest is shown in finding prp factors of (a^nb^n)/(ab) other than the Mersenne Cofactors, I would strongly appreciate more and more searches for these factors from others. I am working hard at prp testing factors generalized base repunit forms, and a^nb^n where a = b+1 more specifically. Thanks if someone knows how to do the trial division of these forms.
