From:	"D. Vrabel" <dv207 AT hermes DOT cam DOT ac DOT uk>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Longer than double?
Date:	Fri, 16 Oct 1998 15:25:37 +0100
Organization:	University of Cambridge, England
Lines:	18
Message-ID:	<Pine.HPP.3.96L.981016151736.21533D-100000@tw100.eng.cam.ac.uk>
References:	<706vqe$op9$1 AT news DOT interlog DOT com>
NNTP-Posting-Host:	tw100.eng.cam.ac.uk
Mime-Version:	1.0
X-Sender:	dv207 AT tw100 DOT eng DOT cam DOT ac DOT uk
In-Reply-To:	<706vqe$op9$1@news.interlog.com>
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

On Fri, 16 Oct 1998, Guruman wrote:
>     I am doing some programming that heavily utilizes floating-point math,
> using DJGPP. This code would benefit from having variables longer than the
> precision that 'double' offers. (some juicy chaos theory stuff). The problem
> is, I've reached the precision limit of my variables, and my plots look like
> enormous Atari 2600 pixels, even at 1024 x 768.
>
>     Are there any add-on packages (or hacks) for DJGPP that provide
> extra-large precision functionality? I have seen products of this nature for
> other compilers, but wouldn't know where to begin modifying them to support
> DJGPP.
You can use long double which uses 10 bytes and not 8.  Or there are
also arbitrary precision maths libraries. One I have heard about was used
in the Fractint fractal program (try looking at
http://spanky.triumf.ca/www/fractint/fractint.html)

David.  

- Raw text -