float lower limit at given precision

Jacopo Nespolo j.nespolo at gmail.com
Tue Feb 28 19:33:42 CET 2012


On 28 February 2012 18:01, Zimmermann Paul <Paul.Zimmermann at loria.fr> wrote:
>
> I guess what you need is 2^(-prec), where prec is the (binary) precision of
> your mpf_t numbers.

Yes, thanks a lot!
Also thanks to everyone else who answered. Following the precious
suggestions I got, I found that the best solution for my problem is to
use the MPFR wrapper for C++ by Pavel Holoborodko
(http://www.holoborodko.com/pavel/mpfr/) together with Eigen for the
linear algebra part.

So far they seem to be working together quite smoothly, though my
results differ from my rapporteur's by a few parts in 10^16, which
happen to be of the order of the double precision epsilon. He got his
results through Mathematica.

If I find anything worth reporting to the community I'll open a new thread.
Thanks again so much for your help.
J


More information about the gmp-discuss mailing list