[Insight-users] Hashing 3D euclidean vectors?

[motes motes]

I need to hash regular spaced points located in 3D such that points
that are close to each other hash to the same value (gets stored in
the same chain) and points that are far from each other hash to
different values (are stored in different chains). This way I can
given a random point find the chain with the closest neighbors in
constant time.

Can someone recommend a good hash function that does the above?