Author: Robert Hyatt
Date: 08:24:43 12/23/03
Go up one level in this thread
On December 23, 2003 at 00:52:37, KarinsDad wrote: >On December 23, 2003 at 00:15:44, Robert Hyatt wrote: > >>On December 22, 2003 at 23:39:08, KarinsDad wrote: >> >>>What structure do people use in their programs to store a position? >> >> >>A chess board -> a position? >> >>12 bit boards, one for each piece type. > >Sorry for being dense, but I do not quite follow this. > >So, each piece bit board is the normal 64 bits (except the pawns which are 48 >bits) or 8 bytes (8 bit byte machine) * 10 piece types + 6 bytes * 2 piece types >= 92 bytes? Yes, except I don't do 48 bit things for pawns.. I use 64 bits for everything to give uniform square naming. > >Isn't there additional information needed like side to move, rook castle >possibility, and e.p. pawn? I guess if you stored the pawns as normal 64 >bitboards, you could fill in the extra 8 bits on either end with this data, but >then you would have 96 bytes. Or, you could just add another two bytes for 94 >bytes. It seems like a lot of data per position though (if you have data >persistence). Again, when you say "position" what do you mean? I am talking about the internal chess board representation. If you are talking about (say) book positions, the answer would be different. Hash table positions would be different. Etc... > > >>An array of 64 elements (or more) one for each square. > >Each element would need the 4 bit piece (e.p. pawn, castle rook, side to move >king, empty, and white and black for 6 piece types). 64 elements * 4 bits = 32 >bytes I don't quite follow. You need _one_ side to move bit, wtm=1, btm=0, for the _entire_ position, not for each square. It _really_ sounds like we are talking about two different kinds of "positions"... > >Is this correct? Or is side to move known from even or odd depth in the graph >and that "piece" is replaced with something else? > > >>In theory you could define a position as a 32 element array, one for each >>possible piece, telling what the piece is and what square it is on. But this >>has not been used in a _long_ time so far as I know, as it is too slow to deal >>with. >> >>I'm sure you could create others. But probably none would be any better than >>one of the first two I mentioned.] > > >Why are these better than other methods? Speed to access them...
This page took 0 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.