[SciPy-Dev] Is odeint re-entrant?

[Benny Malengier]

2014-04-02 7:46 GMT+02:00 Daniel Farrell <boyfarrell at gmail.com>:

> Dear list,
>
> The docs warn that the lsoda, vode, zvode solvers in scipy.integrate.ode
> are not re-entrant. How about the the implementation for
> scipy.integrate.odeint? The docs state that is used the lsoda solver, so I
> guess that it is not re-entrant.
>
> To check my understanding, re-entrant prevents me from running multiple
> instances of the above solvers on different processes?
>

It would be strange if different processes were not possible, I think it's
more a question of those solvers storing internal state, so that you can
call step to advance one step, and still use BDF. So, it should only be in
a single process. Starting a new problem calls initiate and resets the
internal state of the solver.

The modern cvode does not have this warning, so you can certainly use that
to instantiate different problems at the same time in a thread, eg via my
code https://github.com/bmcage/odes.

Benny


>
> Best wishes,
>
> Dan
>
>
> _______________________________________________
> SciPy-Dev mailing list
> SciPy-Dev at scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scipy-dev/attachments/20140402/dc309a19/attachment.html>