Hey anon,

I know that Krylov solvers preconditioned by algebraic multigrid are expected to converge in number of iterations independent of problem size. This is definitely true for elliptic problems. Is it also true for parabolic problems?