The variant of Robin Hood hash table isn't a double hashing scheme like the original research paper. This is more like linear probing with a collision resoluti

The variant of Robin Hood hash table isn't a double hashing scheme like the original research paper. This is more like linear probing with a collision resolution trick to reduce worst case in lookups. The article suggests using quadratic probing instead, but that has two weaknesses. Cache locality is worse than linear probing. But the deal breaker is that the hash table must have a prime number size for correctness. This makes the hash table grow quite quickly once the size reaches tens of thousands. And finding primes is another issue.