What package to use.

[Paul Zimmermann]

       Dear Joel,

> Date: Tue, 16 Sep 2008 18:52:24 +0200
> From: "Joel Kronander" <joekr552 at student.liu.se>
> 
> Hi all
> 
> I am a new user of the GMP package. I am doing a scientific project ("high
> quality volume rendering") and I am in need of
> a good package for arbitary precision floating point operations and
> representations. My general concern is more with accuracy then speed.

in that case I would recommend MPFR (www.mpfr.org). Other floating-point
packages with accuracy guarantees are NTL's RR class in binary, Maple and
decNumber in radix 10 (although only for the 4 basic operations in Maple).
As for speed, you'll find a detailed comparison on
<http://mpfr.org/mpfr-2.3.0/timings.html>.

> Wich package do you think suits my needs the best, GMP or something else?
> 
> /best regards
> Joel Kronander

Depending on your exact needs, you can also use (arbitary precision) interval
arithmetic libraries, like for example MPFI, which is based on MPFR.

Best regards,
Paul Zimmermann