help-gsl
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Help-gsl] Possible bug in gsl_odeiv2_step_bsimp


From: ali franco
Subject: Re: [Help-gsl] Possible bug in gsl_odeiv2_step_bsimp
Date: Sat, 15 Oct 2011 14:32:29 +1300

I am guessing that there is some memory leak in bsimp. something should be
freed but is not.

On Sat, Oct 15, 2011 at 1:58 PM, ali franco <address@hidden> wrote:

> Dear all,
>
> I have a parameter defined as
>
> struct parameter{ double wave_num;
>                           bool  partial;
>                           int count;
>                          };
>
> I pass this parameter and the Jacobian to gsl_odeiv2_system and every thing
> works fine the first time around. But if I immediately call the solver again
> the integration fails. I am pretty sure it is a bug in step_bsimp because
> the error does not appear with other solvers. For example
> gsl_odeiv2_step_msbdf or for that matter even if I use gsl_odeiv2_step_rk8pd
> (which is not really recommended for stiff problems) every thing works just
> fine.
>
> So I am reporting that with a parameter defined like above and on the
> second and third calls to gsl_odeiv2_step_bsimp it fails.
> gsl_odeiv2_step_msbdf works just fine, which lead me to conclude that bsimp
> has a bug.
>
> Can someone please say something about which is the preffered stiff solver?
> bsimp or msbdf?
>
>


reply via email to

[Prev in Thread] Current Thread [Next in Thread]