How to optimize the quadratic sieve setup arguments?

Question

What would be the best strategy optimizing the quadratic sieve setup?

Below is the list of the configurations used by my quadratic sieve algorithm, I would like to build a program that will run several tests to find the optimal setup, but the problem is that there is a non-trivial dependency between the params and I simply don't know how to find it. Please help me.

• primeBaseSize - The number of primes in the prime base
• maxPrimeThreshold - Is used to calculate the logs threshold
• minPrimeSize - Used to limit the sieving to start from primes higher than this value
• loopsSize, loopsCount - loppSize times loopsCount is equal to the sieving bound
• longRoundingError - A threshold for picking the max long value for the vector extraction phase

Answer 1

Get a copy of Prime Numbers - A Computational Perspective by Crandall and Pomerance. See here for a review. It is rather old (2005), but then so is the Quadratic Sieve. It has an extensive discussion of the algorithm, by its inventor Carl Pomerance, and tells you everything you need to know.