mpn_sqrtrem{1,2} - patch for pure C implem

[Marco Bodrato]

Ciao,

Il Mer, 7 Giugno 2017 1:40 am, Torbjörn Granlund ha scritto:
> "Marco Bodrato" <bodrato at mail.dm.unipi.it> writes:
>   For some computations, even for the 64-bits version, a 32x32->32
>   multiplication is enough. I #define-d the gmp_uint_fast32_t type,
>   but currently it is not really guaranteed "fast"...
>
> Why do you want that?
>
> I believe that would be an improvement on very few platforms.

Because it can be an improvement on some platforms :-) even if only on a
few of them :-)
Nevertheless I'll #define gmp_uint_fast32_t mp_limb_t

> Of course, if it is used for 2-limb sqrt on 32-bit systems, it's a
> good improvement.

It is used.

-- 
http://bodrato.it/