Macro is kinda funny, the stuff that can be done with algebra, doesn't really need anything more, and the stuff that does need more, is a whole different ballgame, mathematically speaking - basically everything in grad macro requires dynamic optimization, no way around it. One of the problems with Romer's book is that it teases you with all these results, but you can't actually derive them unless you know how to do dynamic optimization (which the book doesn't teach you).
That said, I found "Macroeconomic Theory:A Dynamic General Equilibrium Approach" by Wickens to be a decent "roadmap" to graduate-level macro.