[petsc-dev] TS monitoring at fixed times?

[Jed Brown]

On Sun, Oct 23, 2011 at 14:26, Barry Smith <bsmith at mcs.anl.gov> wrote:

>    It would be nice to be able to run the TS monitors so that they are
> called at fixed "times" instead of just fixed "timesteps". For example, save
> to a file the solution at each .1 seconds.  This would be useful for doing
> grid convergence studies, parameter studies, etc.
>
>     Presumably this could be done by having a TS monitor check if it has
> passed the next "time-point" and then interpolate the solution back (or
> something back) to that point and then (somehow) call the regular TS monitor
> that plots the solution, saves it to a file, etc.
>

Yes, this is trivial to implement with a custom monitor (just have it call
TSInterpolate). For an automatic monitor, we need an API for specifying
which times are desired.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111023/91d9b226/attachment.html>