Extending the mpn interface

[Niels Möller]

Bill Allombert <Bill.Allombert at math.u-bordeaux1.fr> writes:

> Yes, please. What I would really like for PARI/GP:
> mpn_powm, mpn_divexact, mpn_rootrem

Those should be fairly straightforward.

> mpn_halfgcd (this is implemented as part of subquadratic gcd)

Here, I think we may need to spend some effort on the interface. Have
you looked at the currently internal hgcd functions?

What use do you have for half gcd? The two applications I'm aware of are
subquadratic gcd/gcdext, and "rational reconstruction" (related to
Axel Thue's Lemma).

> a way to precondition the modular reduction (this is implemented as part
> of mpz_powm).

I'm not entirely sure what you mean, but on the mpn level, I think it
makes sense to make some redc functions public.

Regards,
/Niels

-- 
Niels Möller. PGP-encrypted email is preferred. Keyid C0B98E26.
Internet email is subject to wholesale government surveillance.