We have a system here that uses very large datastructures. For some projects, the size of e.g. a certain hash can reach up to 100_000 elements: The script becomes notably slower when reaching this size. Is there any way to speed this up? I was thinking of using tie with a suitable Tie:: module that is faster than perl's default implementation. I guess it should be possible to be much faster if you limit yourself to storing literal values only, not references. Any recommendations?