On Mon, Sep 8, 2008 at 11:45 AM, Olivier Teytaud <[EMAIL PROTECTED]> wrote: >> >> By my recent experiments, 8~9 * (threads - 1) ELO is lost. This >> matches my earlier result well. > > > Do you have tricks for avoiding redundancies between simulations ? > I suggest simple tricks like "do not go to node X if there is a thread > currently in node X" > (simply by setting the score of the moves leading to node X to zero until > you get out of the node).
That doesn't sound like a good idea to me, since it would force most threads out of the interesting moves at the root. If you are using UCT, you can simply increment the number of plays when a thread starts exploring this branch; the number of victories will be incremented later, if it needs to be incremented. This way other threads will be discouraged from taking this move, but not disproportionally. By the way, are people aware that you can do atomic increments with one line of assembly code? Álvaro. _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
