Xref: news2.mv.net comp.os.msdos.djgpp:2071
Newsgroups: comp.os.msdos.djgpp
Subject: Re: floating point error
Message-ID: <31500581.3256@vnet.ibm.com>
From: Jim Groves <jgroves AT vnet DOT ibm DOT com>
Date: Wed, 20 Mar 1996 08:17:53 -0500
References: <4iolb6$6ph AT brachio DOT zrz DOT TU-Berlin DOT DE>
Organization: Celestica
NNTP-Posting-Host: a.socks2.server.ibm.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Lines: 18
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

Wolfgang Winkler wrote:
> 
> Hi,
> I use some mathematical source-files (lp_solve) and they work fine when
> compiled by gcc 2.7.0 on sun and also when compiled by djgpp V1.x (gcc 2.6.3)
> but the program chrashes sometimes (floating point eception) when it is
> compiled by
> djgpp V2 (gcc 2.7.2). The computer is running with win95.
> 

I had a similar problem with floating point exceptions with an
electrical network analysis program (spice). After reading earlier
threads concerning libm, I put a floating point reset (fpreset?)
as one of the first executable lines in the program, and the
exceptions *seem* to have stopped. Sorry for the fuzziness about
fpreset, I don't have djgpp installed here at the office.

Jim Groves  jgroves AT vnet DOT ibm DOT com