Talk:N-queens minimum and knights and bishops: Difference between revisions
Talk:N-queens minimum and knights and bishops (view source)
Revision as of 15:13, 11 May 2022
, 2 years ago→Performance note: <del>'d my cheat remark
m (→Performance note: <del>'d my cheat remark) |
|||
Line 20:
== Performance note ==
One thing I have found so far is that <code>integer m=1; while not solveable(m) do m+=1 end while</code> is at least five times faster than finding "any solution" and then exhaustively eliminating the existence of anything better. Since it takes my current approach(/that^) around 15 mins (3 mins with cheat below) to solve, I'm currently writing a GUI version so you can at least explore (eg) the 8x8 solutions while it is still cranking on with the 10x10s in the background. <del>If you (cheat) and set m to the right answer to start with it is five times faster again. I might yet go hybrid, as in cheat first, then prove.</del> --[[User:Petelomax|Pete Lomax]] ([[User talk:Petelomax|talk]]) 01:03, 24 April 2022 (UTC)
:As usual with Nigel's tasks, the difficulty is not so much finding a way to do them but finding a reasonably quick and efficient way. If he's having to resort to sophisticated stuff like the now deprecated Microsoft.SolverFoundation for his own F# solution, then I suppose that's telling you something.
| |||