Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.graphics.apps.gnuplot > #367

Re: trying to get a fit to work.

Followups directed to: comp.graphics.apps.gnuplot

Show all headers | View raw

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. 

Back to comp.graphics.apps.gnuplot | Previous | Next — Previous in thread | Find similar

Thread

trying to get a fit to work. aviphysics <aviphysics@gmail.com> - 2011-06-07 16:32 -0700
  Re: trying to get a fit to work. Hans-Bernhard Bröker <HBBroeker@t-online.de> - 2011-06-08 23:02 +0200
  Re: trying to get a fit to work. sfeam <sfeam@users.sourceforge.net> - 2011-06-08 21:48 -0700
    Re: trying to get a fit to work. Hans-Bernhard Bröker <HBBroeker@t-online.de> - 2011-06-09 22:26 +0200
      Re: trying to get a fit to work. sfeam <sfeam@users.sourceforge.net> - 2011-06-09 18:29 -0700

csiph-web