Author: Uri Blass
Date: 02:35:01 04/03/02
Go up one level in this thread
On April 03, 2002 at 04:43:10, Slater Wold wrote: >On April 03, 2002 at 03:52:25, Tom Kerrigan wrote: > >>On April 02, 2002 at 13:22:53, Slater Wold wrote: >> >>>I am going on vacation tomorrow for 2 weeks. When I get back I will profile >>>Crafty, and then look into the possibility of porting the heaviest parts into HW >>>generated via FPGA's. It is very possible that I will stick with making the >>>move generator first, and then move on to the harder parts. >> >>I don't understand the Crafty fetish. > >It's just the program I know the best (which doesn't say much). In the begining >this was just going to be a move generator, but I am not convinced that is worth >the effort. > >> >>A hardware implementation of any part of Crafty _should_ be completely different >>from its software implementation. As long as you're doing something completely >>different, why try to maintain consistency with something arbitrary? > >After reading for the last 12 hours, I will more than likely take the Belle/DT >route. Which is not bad. We never really got to see how those things >performed. > >>Also, what kind of FPGA are you planning to use, that you're going to implement >>an entire eval function? I ask because I believe it would be impossible to fit a >>single set (6) of piece-square tables on the largest Virtex, not to mention >>actual logic, although I do notice that the Virtex-II Pro has 22592 slices >>(!!!). I also bet it costs more than my life's savings... > >I wasn't *planning* on using a very large one. But since I am going to be >adding eval, then I will of course have to. Another option is to trying to change the evaluation function of Crafty to do it faster. Crafty is using evaluation function only in the leaves. It may be possible to try to calculate the full evaluation one time in the root and evaluate only the change in the evaluation after every move. If evaluating the change in the evaluation can be done clearly faster than it means that it is going to do crafty's evaluation clearly faster. I do not think that it is easy to do it and one disadvantage of this idea may be that changing the evaluation function is going to be less convinient. There may be more problems that I do not see now. Uri
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.