2009/8/25 René van de Veerdonk <rene.vandeveerd...@gmail.com>: > Nonetheless, bitmap-go as a topic keeps resurfacing on this > mailing list every once in a while and nobody ever put solid data and a > reference implementation behind it. That is what I wanted to accomplish with > my mockup.
I think this is interesting information, so I think it's good you tried it. It basically means unless someone has a much smarter idea how to implement bit-sets, the performance is not enough to go through the trouble for most programmers. > So, it looks that I was overly pessimistic in terms of performance drop per > move, which is a factor of 2.4x (with little effort to optimize for 19x19). > But, with 4.1x more moves, this still resulted in a 10x speed penalty. With > the provided reference numbers, Libego only drops 4.5-5.0x, indicating that > there is almost no performance loss per move (1.2x). Is this difference > roughly in line with others expectation? Yes, I'd expect that. The main speed-determining factor in traditional implementations is merging. That depends on the average length of the merge. Obviously this can be much longer on 19x19 as a worst case, but on average I wouldn't expect it to make much difference compared to other board-sizes. Mark _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
