>>12990508It's easiest just to use a lot of fast sieve methods, then whatever comes out use a primality test on. Really there can't a "singular" primality test which is best, but combining many smaller tests with RKS has the best time-space tradeoff.
