18 Sep
2006
18 Sep
'06
5:37 p.m.
On 9/18/06, Joaquín Mª López Muñoz <joaquin@tid.es> wrote:
As it happens, I did in the past a little sketch of how a dynamic multi_index_container (no intrusiveness) could be implemented, and basically all of the above coincides with my stuff, except one thing: what is this MPHF thing and what do you use it for?
It is used to lookup the appropriate sub_node in super_node in constant time since the offset isn't always the same. I use a minimal perfect hash function so that the sub_node vector doesn't waste any space. Its size always equals the number of contained sub_nodes. Kevin