num::assemblable_grid_operator

Defined in header <pde/concepts.hpp>

template <class Op>
concept assemblable_grid_operator = linear_operator<Op> && requires(const Op &A) {
  { A.to_sparse() };
};

Grid operator that can also materialize itself as a sparse matrix.

Krylov methods need only the action. A direct solve needs the matrix. An operator satisfying this supports both, so the choice of solver does not change how the discretization is written.