New code for primality testing

[Torbjörn Granlund]

"Marco Bodrato" <bodrato at mail.dm.unipi.it> writes:

  If we can verify the source of the second claim, since we just implemented
  BPSW, we should decide to return 2 (surely prime) for all "probable"
  primes under this threshold.

We could check ourselves to a gradually larger bound, then insert our
latest determined bound into the current GMP code and return 2 for
smaller primes.

-- 
Torbjörn
Please encrypt, key id 0xC8601622