Recent changes to mpn_get_str/mpn_set_str
bodrato at mail.dm.unipi.it
Fri Feb 3 18:50:58 UTC 2017
Il Gio, 2 Febbraio 2017 5:02 pm, Torbjörn Granlund ha scritto:
> "Marco Bodrato" <bodrato at mail.dm.unipi.it> writes:
> Maybe I'll try, the basecase...
> I take this as a promise. :-)
> If coming from the sub-quadratic code, there will be a suitable
> If coming directly from the external caller (e.g. mpz_get_str) no such
> power will be available. One should be computed with mpn_pow_1 (or a
... and for the basecase I mean in particular this scenario.
> GET_STR_DC_THRESHOLD for base 10. For CPUs with mpn_mulhi (currently
> none) tabling also the inverses (or just the inverses!) also makes sense.
The mulhi issue is a big one... not only we do not have it specialised for
any CPU, but we don't have a pure C version... we did not even decide the
result we want from such a function :-D
More information about the gmp-devel