All,
I have tested out the functionality on a new example and things seem
to be working as desired. This patch is ready to go into trunk. Please
provide any comments/concerns, especially on the memory_solution_history.C
functions.Thanks. On Fri, Nov 16, 2012 at 3:22 PM, Vikram Garg <simulation...@gmail.com>wrote: > Hey all, > This is the first and hopefully hardest step towards > getting time-dependent adjoints support going. The patch attached will > enable the user to save and retrieve solutions at all timesteps. The user > will > do this by declaring a MemorySolutionHistory object and then passing a > reference to that object to a SolutionHistory object via a setter > function. The library will then automatically store the solution at > all timesteps in an appropriate data structure. > > The next step is defining a new adjoint_advance_timestep function that > will use the retrieve function in MemorySolutionHistory and prepare > the vectors for reassembly of the stiffness matrix and doing the > adjoint.solve(). Please provide your comments/concerns/corrections. > > Before being able to compile with the patch, you will need to update > your simlinks by navigating to /include/libmesh and doing > ./update_links.sh > > Thanks, > Vikram. > -- Vikram Garg Postdoctoral Associate Center for Computational Engineering Massachusetts Institute of Technology http://www.runforindia.org/runners/vikramg
solution_history.patch
Description: Binary data
------------------------------------------------------------------------------ Keep yourself connected to Go Parallel: DESIGN Expert tips on starting your parallel project right. http://goparallel.sourceforge.net
_______________________________________________ Libmesh-devel mailing list Libmesh-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-devel
