Hi ibd, Guillaume's paper is informative. <http://www.cs.unimaas.nl/g.chaslot/papers/parallelMCTS.pdf>
Don's scalability study is also interesting, though it's not for the number of threads but the playouts for a move. <http://cgos.boardspace.net/study/> (9x9) <http://cgos.boardspace.net/study/13/index.html> (13x13) Hideki Isaac Deutsch: <20090606134839.283...@gmx.net>: >Hi, > >I did some tests with my program about how well it does using 2 threads >instead of using only 1 thread. I played 200 games of 9x9 with 5 min SD using >gogui's twogtp. Here's the results: > >rango (my program), 1 thread vs. gnugo: 46.7 +-3.5% wins >rango, 2 threads vs. gnugo: 54.5 +-3.5% wins >rango, 2 threads vs. rango, 1 thread: 54.8 +-3.5% wins > >Do you have any comparable data, also how it scales with more than 2 cores >(which I don't have...)? It seems the difference of an additional core vs. >gnugo is much >bigger than versus the program itself. > >Regards, >-ibd -- g...@nue.ci.i.u-tokyo.ac.jp (Kato) _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
