If you're not using FEMSystem subclasses with formulations whose implementations require overriding mass_residual (e.g. SUPG stabilization), this change doesn't affect you and you can disregard this message.
If you are using such implementations, your name is probably John Peterson, we've already discussed this by private email, and you can disregard this message. Just in case there are more affected users out there (or in case anyone has suggestions or questions, or just because I like to hear myself talk): Currently, FEMSystem::mass_residual() is evaluated with elem_solution equal to the start or the end values at a timestep, and those two evaluations are finite differenced to get the time derivative terms in a formulation. That works great based on the assumption that the start and end values are integrated on elements of the same size. Mesh motion can cause elements to expand or contract, breaking that assumption. To properly do time integration on moving meshes, I'll be changing the TimeSolver implementations to call mass_residual() with elem_solution equal to the difference between two timesteps' values. For most codes (which can just use the default FEMSystem implementation of mass_residual) no changes are necessary. For codes with only linear stabilization terms (like our shallow water code, John; it looks like I was wrong when said before that I thought it was vulnerable) no changes are necessary. Even codes with nonlinear stabilization terms which are properly written (to use elem_fixed_solution rather than elem_solution for the parts in test functions) should continue to work correctly. But: any incorrectly written nonlinear stabilization terms will change from "just slightly inconsistent" to "unstabilized or horribly broken". Try to think of that as a good opportunity to fix them. --- Roy ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Libmesh-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-devel
