Path: csiph.com!x330-a1.tempe.blueboxinc.net!newsfeed.hal-mli.net!feeder1.hal-mli.net!feeder.news-service.com!85.214.198.2.MISMATCH!eternal-september.org!feeder.eternal-september.org!.POSTED!not-for-mail From: sfeam Newsgroups: comp.graphics.apps.gnuplot Subject: Re: trying to get a fit to work. Followup-To: comp.graphics.apps.gnuplot Date: Thu, 09 Jun 2011 18:29:28 -0700 Organization: gnuplot development team Lines: 20 Message-ID: References: <62b76acf-a50c-49f2-a02a-908398651e5b@q12g2000prb.googlegroups.com> <95coieFepcU1@mid.dfncis.de> Reply-To: sfeam@users.sourceforge.net Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 8Bit Injection-Date: Fri, 10 Jun 2011 01:29:30 +0000 (UTC) Injection-Info: mx04.eternal-september.org; posting-host="joalffpZlyHxFtpFd/PRrw"; logging-data="20898"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+YNLcuk9EqPMcy+UH/xmiJ" User-Agent: KNode/4.4.9 Cancel-Lock: sha1:iRJX6YViX6CzGsiit/gu+6s/4ew= Xref: x330-a1.tempe.blueboxinc.net comp.graphics.apps.gnuplot:367 Hans-Bernhard Bröker wrote: > On 09.06.2011 06:48, sfeam wrote: >> I am guessing that you are using a machine that has configured in the >> internal version, and there is an overflow bug in it somewhere. > > That's no bug, it's a deliberate, reasonably error reaction. It's a > bona fide floating point exception being raised by the implementation, > which under the usual roles of standard C library routines is apparently > _supposed_ to send out a floating point exception. That's the same > reason implementations started to return "undefined" for > exp(-largenumber), one after the other, a while ago. The bug is that the algorithm should not invoke exp(something-that-will-overflow) at all, since the erfc() function has already reached its limiting value of 0. I see no point in reporting that erfc(x) = 0 for x=28, which is roughly the limiting precision, and then returning something other than 0 for x>28.