This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: gsl_siman bug
- From: Brian Gough <bjg at network-theory dot co dot uk>
- To: "Peter S. Christopher" <peterc at midway dot uchicago dot edu>
- Cc: gsl-discuss at sources dot redhat dot com
- Date: Sun, 14 Jul 2002 21:44:28 +0100 (BST)
- Subject: Re: gsl_siman bug
- References: <Pine.GSO.4.21.0207111608500.27892-100000@harper.uchicago.edu>
Peter S. Christopher writes:
> I think I've found a bug in the gsl_siman_solve routine. I'm using
> gsl 1.1.1. The bug resides in the file gsl-1.1.1/siman.c on line #99. It
> currently reads
>
> } else if (exp (-(E - new_E)/(params.k * T))
> * gsl_rng_uniform(r) < 0.5) {
>
> I believe it should read something more like
>
> } else if (exp ((E - new_E)/(params.k * T)) >
> gsl_rng_uniform(r)) {
>
> The first version isn't consistent with the documentation. Further, and I
> think that this is important, it doesn't allow users to set the
> temperature such that he/she starts out with near 100% acceptance. Let me
> know if there's any other info I can provide.
Thanks for the bug report.
regards
Brian Gough