Author: KarinsDad
Date: 21:43:57 02/19/99
Go up one level in this thread
On February 19, 1999 at 20:40:34, James Robertson wrote: >On February 19, 1999 at 17:20:55, KarinsDad wrote: > >>On February 19, 1999 at 16:30:16, Will Singleton wrote: >> >>>On February 19, 1999 at 15:07:08, James Robertson wrote: >>> >>>>On February 18, 1999 at 21:09:33, Will Singleton wrote: >>>> >>>>> >>>>>On February 18, 1999 at 20:21:25, James Robertson wrote: >>>>> >>>>>>Yes! >>>>>>Finally they gave my a FICS account. It has the rather boring name of "JRCP", >>>>>>but John Perry said I could change it when I think of a name for my program. >>>>>>Anyway, it'll be glad to play any chess programmers here (and their programs >>>>>>too, of course ;). >>>>>> >>>>>>James >>>>> >>>>> >>>>>About time. :) I'll go on fics tonight. >>>>> >>>>>Will >>>> >>>>Thanks for the games last night! Amateur really crushed me. Fortunately, I found >>>>several huge bugs that trashed my program's play. Among others, it would do >>>>futility and extended futility pruning only when there were extensions! That is >>>>how it missed that simple mate and played Bc5??. Also, I fixed the pawn hash; >>>>only some of the positions recieved pawn scores. I'm ready to play you again! >>>> >>>>James >>> >>> >>>It sounds like you have a lot of advanced stuff in there. How long have you >>>been working on this? It took me well over a year to get the pawn-hashing in, >>>let alone extended futility pruning. I don't use that at all, if you're talking >>>about the DarkThought stuff. Heck, I didn't even know how to properly use >>>extensions until recently (probably still don't). >>> >>>Do you also use rotated bitboards and book learning? :) >>> >>>I did notice that you're king safety was weak, I was able to get in there pretty >>>easy. But otherwise it played very well. >>> >>>Will >> >>Do either of you have your games that you played against each other? > >I don't think so :(. > >> >>I wanted to use some games from amateur programs playing each other (as well as >>from the stronger programs) to use as test cases for my program (if it ever >gets >>done). > >Well.... don't embarass me by posting them. :) > >> >>Thanks :) >> >>KarinsDad >> >>PS. My program uses rotated bitboards, but I'm nowhere near futility pruning or >>null moves. > >Rotated bitboards are awesome. How many nodes per second does your engine get? >Currently I get 100k to 200k nodes in an average middlegame position on my P233. >Endgame goes way up. With just a simple evaluation function and without all the >pruning and stuff, I get often 300k nps in the middlegame (it doesn't search >very deep, though). > >James The last I looked, it was about 96K on a 400 Mhz, but that was with some major evaluation yanked out, so I'm sure it's much slower (maybe 30-40K). However, I do not have the hash table working yet, so I'm sure that it will pick things up a lot because I am still doing multiple mallocs, arghhhhh, and am waiting to do one malloc per major data structure and my own memory management when I do the hash work. The real problem is that I have guests for the next two weeks, so I will not be able to spend a second on it (although I will be going skiing for the next week, so that should make up for it, and of course you guys won't have to listen to my ramblings ;) ). KarinsDad :) :)
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.