Newsgroups: comp.os.msdos.djgpp
From: Elliott Oti <oti AT phys DOT uu DOT nl>
Subject: Re: high precision nonlinear math functions ?
Sender: usenet AT phys DOT uu DOT nl (News system Tijgertje)
Message-ID: <Pine.OSF.4.03.9811131120560.27881-100000@ruunf0.phys.uu.nl>
In-Reply-To: <4.1.19981112171650.00a26d70@hal.nt.tuwien.ac.at>
Date: Fri, 13 Nov 1998 10:30:56 GMT
X-Nntp-Posting-Host: ruunf0.phys.uu.nl
Content-Type: TEXT/PLAIN; charset=US-ASCII
References: <4 DOT 1 DOT 19981112171650 DOT 00a26d70 AT hal DOT nt DOT tuwien DOT ac DOT at>
Mime-Version: 1.0
Organization: Physics and Astronomy, University of Utrecht, The Netherlands
Lines: 44
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

On Thu, 12 Nov 1998, Anton Helm wrote:

> I've some problems with numerical precision 
> of some math functions.
> 
> Is there any math lib available that can do these
> computations in long double ?
> 
> I would need e.g.:
> 
> log(tan(b/2.0 + PI/4.0))
> 
> asin(e * sin(b))
> 
> atan(exp(f))
> 
> Unfortunately all of them are inside an iteration loop
> so they really need to be fast.
> 
> I recently got an infinite iteration loop due to low precision
> (all other computations are done in long double).
> If I change the break criteria of the iteration loop so that I
> jump out before the error occurs the results are not precise enough :-(


Hmm. If you're looking for double precision  math software you might want
to search the Guide to Available Mathematical Software at
http://www.nist.gov/gams/
for free and commercial software and libraries.

Netlib, at http://www.netlib.org is also a good starting point.

But if you are planning to use long double precision transcendental
functions in an inner loop, then you can't expect much speed anyway. And
if your source code is suffering from numerical precision problems
because it's operating at the limits of i86 double resolution, you should
seriously consider re-examining the algorithms you are using. 

Cheers

  Elliott Oti
  http://www.fys.ruu.nl/~oti
  Eh? Where's my sig?