Martin Kronbichler wrote: >> Why can not make a code "MatrixTools::apply_boundary_values function" for >> PETScWrappers::MPI::BlockSparseMatrix data type? Why is that difficult to >> make this code? > > I don't think it is difficult. Such a function is implemented for the > class Trilinos::BlockSparseMatrix, and it should work just the same for > PETSc as well. I think the only reason why there is no such function yet > is that nobody has needed it... So you might have a look in > deal.II/source/numerics/matrices.all_dimensions.cc to find out how you > could implement that for PETSc. >
I already did this "work" and committed it to the svn a few minutes ago. bye, Timo _______________________________________________
