GeorgeM

distribution is almost certainly the name of the game if we want to get much beyond n = 16. just to confirm when you say "database" do you mean off...

for context, when i tried to do n=14 it did this:  and that was just trying to store the n=14 cubes. yes we should partition data based on cube...

> if we can come up with a faster way to rotate the cubes, or a fast orientation independent hash function then yes. either would make an excellent first pass...

this is cool and does seem to reduce the number of comparisons needed, unfortunately comparisons are fast, and the math you are doing to work out which to remove is...