> > Against a standard opponent (not the same program), compare scaling in time > vs scaling in cores. >
Hi. We've published this kind of numbers here: http://hal.inria.fr/inria-00287867/ It's a bit old and with small numbers of cores, but we've submitted recently a paper with bigger scale. Roughly (and not only in Go): - yes, there is some loss when increasing the number of cores, whenever the number of simulations is preserved; I've tried several tricks against this but I've no significant success; - also, the success rate of 2N simulations per move against N simulations per move decreases when N increases; - in particular, there are situations in which chosen moves are bad whenever N is very very very large. Best regards, Olivier
_______________________________________________ Computer-go mailing list [email protected] http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
