Currently, in the control-flow graph view, IGV shows all blocks even if only a small set of nodes is extracted. This has the advantage of providing context (e.g. the user can see the loop nesting level of a certain block she is focusing on) (see small.pdf attached), but makes it hard to localize the extracted nodes in large graphs (see large.pdf attached).

Since both modes have advantages in different situations, a possible solution is to provide a toggle button to show or hide empty blocks in the CFG view.