Interesting. Under the hood FGL is mapping the graph to relatively efficient data structures like Patricia Trees as implemented in Data.IntMap so I would expect it to scale reasonably for inserting edges and mapping over nodes. I agree the memory inefficiency is definitely a limiting factor of the library. As you say I think it is best suited for expressing graph algorithms and if those calculations become the bottle neck a custom solution can be developed with the proof of concept already in place. Out of curiosity what number of nodes/edges would you consider a "moderately sized graph"? My user cases are typically on the order of 500-1000 nodes with a similar number of edges that require bfs and topological sorting.