Dear all, with the latest commit 2a5d912dc940... the polynomial factorization now uses the cln class cl_UP_MI instead of the ad-hoc invented UniPoly class. Maybe it is interesting to note that my benchmarks show no improvements in speed at all (rather some tiny change for the worse). This is kind of disappointing, because I expected some decent improvement due to the inefficient and times stupid ways UniPoly was set up. Maybe this recent remark on the cln mailing list gives a good explanation for that finding: Alexei Sheplyakov schrieb:
For example, the same univariate GCD code works about 3 -- 10 times faster (and uses less memory) if I use std::vector<cl_I> instead of cl_UP.
Well, at least the line count in factor.cpp has decreased. Regards, Jens