I spent today working on the Hash Table Shootout project. Check out the updated version on git hub: https://github.com/jimbelton/hash-table-shootout. I added a new fast benchmark that counts the number of unique words in the King James version of the bible (I’ll post about that later), and sped up the existing benchmarks a ton.
Another change was to fix a bug whereby all the C++ hash table implementations were hashing pointer values rather than the actual strings in all of the string benchmarks. This means that the results I previously posted were somewhat incorrect. Here are the updated graphs: