[petsc-dev] TSSolve() shouldn't return a time

[Jed Brown]

On Sun, Oct 23, 2011 at 16:24, Matthew Knepley <knepley at gmail.com> wrote:

> I would definitely make the former the default. Doing a "given amount of
> work" only holds if you turn off adaptivity as well, and if you
> only care about inspecting, I do not see why you would care about a final
> interpolation.
>

It might be more common if people ask for a given number of steps.
(Obviously the cost of that is dependent on the method, a high order RK
scheme will take fewer steps than a multistep scheme, but might cost more to
do it.)

Plenty of other times, people just really don't care about the final time or
they want to make a plot of error per step, in which case they want to see
whatever step sizes the method is using.

I'll change the default and add the interpolated logging interface that
Barry requested this week.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20111023/fa6cedb2/attachment.html>