Re: SoCiS2014 - odepkg

[Jacopo Corno]

Dear Roberto,

I encountered another issue. I am updating other solvers to the new
steppers structure. During some tests I realized that with the current
evaluation of errors using the function 'AbsRel_Norm' the stepsize chosen
for the automatic_stepsize are way smaller than in the previous code where
the comparison was on velta < vtau. Did you encounter the same problem
during your work? Now that I looked more carefully it seems to me that
even ode45 has the same problem...

Thanks, best regards
Jacopo


> Dear Carlo, Roberto, Mattia,
>
> I have a couple of questions regarding the best way to continue my SoCiS
> project: is it better to focus on the functions and solver we want to put
> in the core (that is odeset, odeget, ode45 and ode23s) or it's preferable
> to update all the solvers to the current structure and do a complete new
> release of odepkg as it is?
>
> Secondly a more technical question: in the current version the function
> ode_struct_value_check (that only checks the options) has been created, in
> addition to odepkg_structure_check (that also check if the options are
> compatible with a specified solver). Is there a particular reason for this
> choice? I also have some doubts about the fact that odepkg_structure_check
> always returns errors if an option is wrong whileode_struct_value_check
> sometimes return an error and sometimes just a warning. Is this just a
> bug?
>
> Thanks a lot. With kind regards,
> Jacopo
>
>
> --
> Jacopo Corno M.Sc.
> Technische Universität Darmstadt
> Graduate School of Computational Engineering
> Dolivostraße 15
> 64293 Darmstadt / Germany
>
> Office: S4|10-232
> Phone: +49 6151 16 - 76877
> Email: address@hidden
>
>
>


-- 
Jacopo Corno M.Sc.
Technische Universität Darmstadt
Graduate School of Computational Engineering
Dolivostraße 15
64293 Darmstadt / Germany

Office: S4|10-232
Phone: +49 6151 16 - 76877
Email: address@hidden