Toom-8 testing (mpz level)

Torbjorn Granlund tg at
Thu Oct 8 19:45:56 CEST 2009

nisse at (Niels Möller) writes:

  The region where base case wins looks like bn < THRESHOLD, which is
  nice. I.e., it could use a threshold on the size of the smaller factor
If we implement e.g. toom32 for invariance in the smaller operand, it
seems safe to assume that such a toom function will slowly take over the
basecase space as the larger operand grow.

The largest basecase tongue is for the operand size relation 3:1 i.e.,
6:2.  This means two toom32 could be used.  I bet that entire tongue
would be taken by 2xtoom32 (i.e. toom32 partial in the diagrams) when we
implement invariance.


More information about the gmp-devel mailing list