On Thu, Feb 16, 2012 at 12:58 AM, Daniel Wheeler <[email protected]>wrote:
> > You're quite right! What I said doesn't make any sense. As you say, the > right way is to split it an integrate in time, which seems reasonable. It > remains local that way. > > That's unfortunate - I hoped I was wrong. It turns out for my problem that doing the splitting results in two exponential instabilities that just happn to cancel each other. Doesn't sound like a good numerical recipe! If I don't do the splitting then I have to solve the elliptic problem v_x = f every time step, which is okay in 1D, but for larger systems? Doubt it. So sadly it looks like I won't get to use FiPy this time around. Back to finite difference... Thanks again for all your help.
_______________________________________________ fipy mailing list [email protected] http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
