Semidefinite programs (SDPs) are standard convex problems that frequently found in control and optimization applications. Interior-point methods can solve SDPs polynomial time up to arbitrary accuracy, but scale poorly as the size of matrix variables number constraints increases. To improve scalability, be approximated with lower upper bounds through use structured subsets (e.g., diagonally-dom...