In traditional parallel methods for solving multi-physics or multi-field problems,the system is often first split into several single-field problems and some parallel solvers are then applied to those simpler single-field problems.An outer loop is then applied to carry out the field-by-field iterations.The sequential nature of the outer loop is not desirable for supercomputers with a large number of processor cores.In this talk,we discuss some strategies for the parallelization of the outer loops.