On Tue, 29 Jan 2008, Benjamin Kirk wrote: > We should be able to start simple by scripting up the examples, and > modifying them where appropriate.
Hmm... I've just been working on a "parameterize.pl", which takes a whatever.in file with directives like "refinement_steps = range(1,2,5)" and produces a nested tree of directories with all combinations of those config settings. It's a quick and dirty script that'll take work to be useful for testing, but the idea of scripting the examples to run with many combinations of options is a good one. > ex4 can work with any C1+ basis, C0+ > and we should exercise it like that more frequently. Similarly with > ex10 with the refinement, etc... This is definitely the right way to start. > Along these lines, I'd like to target a new release in a month or so. That should be doable. > Not too much beyond what is currently in the library, but there are > some small mods to reading parallel meshes I'd like to make, and the > EquationSystemsIO backwards compatibility needs to be made > transparent. > > Specifically, I want to be able to use the ParallelMesh to solve a problem > bigger than any single processor's RAM. This includes reading the mesh from > disk, solving a steady linear system, and writing the solution back out. I think this is the right point for 0.7.0. We'll still mark ParallelMesh as experimental and we won't make it default, but once it's useful we can point out it's availability for testing. --- Roy ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Libmesh-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-devel
