Don't use linear probing

I haven't benched this library, but I don't expect it to be very fast compared to other hash map implementations such as Round Robin hashing or Google's Swiss Table.

The issue names no files, tests, or entry points. Start by locating the hash map implementation and benchmarking its current linear probing behavior, then compare the results with the proposed Round Robin hashing or Swiss Table approaches; done requires an agreed strategy and evidence that it improves performance.

Written by the indexing model from the issue text.