Two memory layouts. Data is either stored in a flat array, or with node indirection. Access for unordered_flat_map is extremely fast due to no indirection, but references to elements are not stable. It also causes allocation spikes when the map resizes, and will need plenty of memory for large objects. Node based map has stable references & pointers (NOT iterators! Similar to std::unordered_map) and uses const Key in the pair. It is a bit slower due to indirection. The choice is yours; you can either use robin_hood::unordered_flat_map or robin_hood::unordered_node_map directly. If you use robin_hood::unordered_map It tries to choose the layout that seems appropriate for your data.
Depends on good Hashing. For a really bad hash the performance will not only degrade like in std::unordered_map, the map will simply fail with an std::overflow_error. In practice, when using the standard robin_hood::hash, I have never seen this happening. 041b061a72