Torbjörn Granlund tg at
Wed Sep 4 09:51:18 UTC 2019

nisse at (Niels Möller) writes:

  Good to have some profile numbers! Then microoptimizations elsewhere
  will not make any significant difference.

  To get better speed, we'd need to either increase the size of the prime
  table used for sieving considerably, say by a factor of 2, 5 or 10.

We have several mechanism (perhaps too many) for trial division.  I
haven't followed this thread, I'm just wanting to ensure you are aware
of what we already have:


Please encrypt, key id 0xC8601622

More information about the gmp-devel mailing list