Open JaredCrean2 opened 7 years ago
If we use an approximate factorization preconditioner for solving a steady Newton iteration, we should find a way to reuse it for solving the adjoint.
Yes, the last iteration of Newton may be close enough that the already-factored preconditioner will work well for the adjoint.
If we use an approximate factorization preconditioner for solving a steady Newton iteration, we should find a way to reuse it for solving the adjoint.