Dear all,
I am new to using deal.II and have recently been focusing on learning the mesh_loop framework, which I understand is the recommended modern way to assemble matrices in deal.II. I would like to ask for your advice regarding the integration of multigrid methods as preconditioners when using the mesh_loop framework. I have studied step-74 and step-16, and have been able to implement an IPDG method for solving the Poisson equation using mesh_loop on globally refined meshes. However, I have encountered two key issues where I am uncertain: 1. When using global refinement, is it necessary to call mg.set_edge_flux_matrices(mg_matrix_down, mg_matrix_up); in the solver, or can this step be omitted? 2. For adaptive refinement, I am unsure how to implement the code—specifically how to correctly define the CopyData and CopyFaceData structures and how to assemble mg_matrix_down and mg_matrix_up in the assemble_multigrid function. I have looked into step-39 for guidance, but it uses the older loop function instead of mesh_loop, and I am unsure how to properly translate its approach into the mesh_loop context. I think it is two different ways to loop cells. Best regards, Haicheng -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/dealii/9ca5d038-da90-4a03-b2ac-22620098f35dn%40googlegroups.com.
