This PR manually handles 4-element and 8-element cases (instead of relying on recursion) when creating morton organization, so it reduces the number of recursions and also the invocation of dividing sets. It's another obvious performance improvement.
This PR manually handles 4-element and 8-element cases (instead of relying on recursion) when creating morton organization, so it reduces the number of recursions and also the invocation of dividing sets. It's another obvious performance improvement.