[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-gsl] lmder
From: |
Brian Gough |
Subject: |
Re: [Bug-gsl] lmder |
Date: |
Thu, 22 Jan 2004 19:34:15 +0000 |
Butel Rémi writes:
>
> I discovered with interest the existence of GSL,
> and started to use it for optimisation problems I have to solve. >
> The version used is 1.4
>
> I am sorry, I found 4 bugs which doesnt seem to be in the GSL-database.
> I'm unfortunatly unable to send you the contexts in which they appears,
> but I can send the corrections done to the sources to work around them.
Thanks for the bug reports.
I will take a look at the check for fnorm==0 and non-zero status.
Regarding SIGFPE and overflow maybe you can turn off the exception
with GSL_IEEE_MODE environment variable (see the manual for details)
or rescale your function?
Ideally I'd want to avoid arbitrary limits like 1e+150 in the code.
Checks for +-Inf would be reasonable though.
--
Brian Gough
- [Bug-gsl] lmder, Butel Rémi, 2004/01/20
- Re: [Bug-gsl] lmder,
Brian Gough <=