Author: Russell Reagan
Date: 13:37:40 08/19/03
Go up one level in this thread
On August 19, 2003 at 13:56:53, Dan Andersson wrote: > Dynamically updating attackboards is fairly fast. And you get legal moves and >captures as an added benefit. I'm not sure I understand this. Do you mean you get "legal moves" as opposed to "pseudo-legal moves"? Or do you mean that you get move generation for free? I'm also not sure how you get captures for free. >I favour the attackboard board representation >where each square is an unsigned 32 bit integer. One bit per piece no less. It >can readily be combined with any other scheme AFAIK. > When you make a move you update the attack information of the moved piece and >pieces attacking the to and from squares as well as any captured piece. Castling >and ep are special cases. This might be done with iteration or even using a >bitboard. Whichever is fastest. Like Sune, I'm also a little skeptical that this will be faster, but it wouldn't be the first time I've been wrong :) I am reminded of the time that Christophe said, in regards to bitboards, that while you can compute more complex evaluation terms using bitboards, those complex terms aren't necessary for a strong program, and that any of the necessary things that can be done with bitboards can be done at least as fast with a his representation. So I wonder, are the things you'd be doing with your attack board represnetation necessary? And if so, can it be done at least as fast with another approach?
This page took 0.01 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.