How can I further improve performance in a read-only load?

[wangqiim]

It is a awesome hashmap implement! I have get significant performance improvement(about 2x) after replacing the std:unordered_map (read only from 50,000,000 map).

• Do you have any ideas or suggestions about how can I further improve performance in a read-only workload?

In this case, After once write about 50,000,000 <int64, bytes[280]> and 50,000,000 <bytes[280], int>. I only focus on Read Only performance.

Thanks.

[martinus]

what do you mean by read only, find or iterating? If iteration speed is important you should try https://github.com/martinus/unordered_dense

[wangqiim]

what do you mean by read only, find or iterating? If iteration speed is important you should try https://github.com/martinus/unordered_dense

find