JASory

Apparently the code was still using the old algorithm plus the new one. This should be considerably faster, but still not for the smallest integers. The crossover to beat GNU...

I think the minimum rust version is 1.84, since machine-prime uses const static arrays. Other than that, I think the issue is with code formatting, but I have no idea...

Factorisation of many numbers beyond 2^128, like 2^128+1 or 2^128+25, 2^128+35 fails. But this is an issue from the num-prime library, apparently it just gives up even though pollard-rho will...

> any update? thanks Yes, it's slightly faster now, the code was simplified. Realistically beating GNU factor for tiny integers would require precomputing large prime_inverse tables which I am not...