Recent changes to mpn_get_str/mpn_set_str

[Niels Möller]

tg at gmplib.org (Torbjörn Granlund) writes:

> I am not sure sliding-window powering makes much sense for powering in Z

It could reduce the mumber of mpn_mul_1 calls, but (as always) leave the
number of squarings unchanged.

> (as opposed to in a ring).

But Z *is* a ring ;-)

> For radix conversion, we arrange things so that the bases are never
> small, so the plain mpn_pow_1 should be very hard to beat.

Ah, then sliding window makes less sense, since there will be no
interesting single-limb powers to tabulate. And we probably don't want
so large table entries that mpn_mul wins over repated mul_1.

Regards,
/Niels

-- 
Niels Möller. PGP-encrypted email is preferred. Keyid 368C6677.
Internet email is subject to wholesale government surveillance.