> It shouldn't "crap out", it should reject the step, shorten it, compute a > fresh Jacobian (if the Jacobian was being lagged), and try again. Thanks for > reminding me that this should be done sooner rather than later.
Jed, For my uses this is a very important feature. Two questions: - Will the step size increase after a shorter step has been successful? Will it go immediately back to the maximum size? - Will it be possible to set a minimum step size, below which the TSSolve() or TSStep() will "crap out?" It is obviously important that TS not just chug along without error if a timestep fails. Cheers Rich > > > Isn't this test data from something that is failing? Should it not be there? > > Yeah, was that example converging when you first created it? > > changeset: 15757:896f032fd4a8 > parent: 15750:1f53ec9d4e00 > user: Barry Smith bsmith at mcs.anl.gov > date: Fri Apr 23 15:03:44 2010 -0500 > summary: some code cleanup and further work on DMMG version 2 > > diff --git a/src/ts/examples/tutorials/ex7.c > b/src/ts/examples/tutorials/ex11.c > copy from src/ts/examples/tutorials/ex7.c > copy to src/ts/examples/tutorials/ex11.c > --- a/src/ts/examples/tutorials/ex7.c > +++ b/src/ts/examples/tutorials/ex11.c > [changes] ________________________________ Richard Foa Katz Dept Earth Sciences, Univ Oxford http://foalab.earth.ox.ac.uk