Comment by cyberpunk

Anyone have an idea how it behaves differently from google's jump hash algorithm? The cool thing about google's one is it's so short I can include it in a HN comment:

    int32_t JumpConsistentHash(uint64_t key, int32_t num_buckets) {
      int64_t b = 1, j = 0;
      while (j < num_buckets) {
        b = j;
        key = key * 2862933555777941757ULL + 1;
        j = (b + 1) * (double(1LL << 31) / double((key >> 33) + 1));
      }
      return b;
    }

https://arxiv.org/pdf/1406.2294