Thank you Lawrence, that solved it.


Tomasz


From: firedrake-bounces@imperial.ac.uk <firedrake-bounces@imperial.ac.uk> on behalf of Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk>
Sent: 28 September 2016 18:03:29
To: firedrake@imperial.ac.uk
Subject: Re: [firedrake] FSI coupling disagreement
 

> On 28 Sep 2016, at 17:38, Tomasz Salwa [RPG] <mmtjs@leeds.ac.uk> wrote:
>
> Dear Lawrence,
>
> Yes, I think I solve it directly. I just use numpy.linalg.solve. Could you indicate how to increase the tolerance?


Please see http://firedrakeproject.org/solving-interface.html#specifying-solution-methods

I think you want the ksp_rtol solver parameter.

Lawrence
_______________________________________________
firedrake mailing list
firedrake@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/firedrake