Register the outputs of hierarchical boundaries to contain critical paths within a single module or boundary. Consider registering the inputs also at the hierarchical boundaries. It is always easier to analyze and repair timing paths which lie within a module, rather than a path spanning multiple modules. Any paths that are not registered at hierarchy boundaries should be synthesized with hierarchy rebuilt or flat to allow cross hierarchy optimization. Registering the datapaths at logical boundaries helps to retain traceability (for debug) through the design process because cross hierarchical optimizations are kept to a minimum and logic does not move across modules.