Author: Robert Hyatt
Date: 22:21:29 11/19/02
Go up one level in this thread
On November 19, 2002 at 15:20:33, Gerd Isenberg wrote: > >Hi Bob, > >Yes, but that might also be true for other approaches. > > >> >>Vincent hasn't invested the time, so he is absolutely not "going to get it". He >>wants to >>understand something "right now" rather than having to look at things for months >>to >>"get in the groove". >> >>bitmaps require time to master. he's not going to spend the time, and as a >>result he is >>never going to appreciate them. >> > >We'll see... > >Vincent is a strong Chess Player and i think also good Chess-Programmer. He's definitely a strong chess player. But the "good chess-programmer" is a wild stretch. you only have to consider the reams of disinformation he produces about programming issues that are not just wrong, but badly wrong, to see why I would have that opinion... > >>When I started Crafty, I made myself a 5-year committment to stick with bitmaps. >> Because >>I was pretty sure that when I started, it was going to take a long while to >>start to "think >>bitmaps." I was right. And I wouldn't switch back now for anything. But for >>the first year >>I thought about it several times but stuck it out as I had promised myself I >>would. > >Yes, even in my old Dos-IsiChess i used bitboards, 32 for every possible piece, >incremental updated and an redundant array of 32-bit-piecesets for every square. >A lot of stuff to update during doing/undoing moves. I had even piece-sets for >pieces enprise. During opening and early middlegame where the number of changed >square controls affected by a move was quite less, the performace was fine. But >later on specially in queen or rook endings, the incremental update effort >increased with the number of squares with changed controls affected by a move. > >So i had already some experience in "thinking" bitboards (since 92,93 article >about chess4.x in Levy's Computer Chess Compendum"). I didn't notice it in that book. I first saw it in the 1977 book "chess skill in man and machine" although I waited until 1995 to do a full bitmap program. I had a lot of bitmap stuff in Cray Blitz, but not a full implementation as the vector hardware made the array move generator take almost no time at all. > >I was so fascinated by your "rotated bitboard" article in ICCA-Journal, even if >i read something before from Heinz, if i remember well. That this was the final >release to write a new rotated bitboard engine. One of the early dark thought team members used to communicate with me quite often and I told him about the idea. He immediately started working on an implementation for the dark thought engine. His name was Peter Gilgasch (I am not sure about the last name spelling, it was a 1995 discussion). I had not heard of this in any way prior to developing the idea in a discussion with a couple of graduate students interested in my original intent to rewrite my program using bitmaps back in 1995... > >Regards, >Gerd
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.