Also, I would expect the time settings to be very significant here. Longer time settings should make it easier to see an improvement with increased machines.
On Mon, Aug 15, 2011 at 9:14 PM, Jean-loup Gailly <[email protected]> wrote: > > In our paper we had good results for 19x19 (until 32 or 64 machines as > far as I remember...), but in 9x9 more than 15 machines was nearly useless. > > Good point. I have done all the analysis so far on 19x19. I should do a few > measurements in 9x9 too. > It was probably stupid to use 64 machines in the 9x9 tournament. > > Jean-loup > > > _______________________________________________ > Computer-go mailing list > [email protected] > http://dvandva.org/cgi-bin/mailman/listinfo/computer-go >
_______________________________________________ Computer-go mailing list [email protected] http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
