Author: Dann Corbit
Date: 12:27:20 05/14/01
Go up one level in this thread
On May 14, 2001 at 15:01:10, Heiner Marxen wrote: >On May 14, 2001 at 13:50:51, leonid wrote: > >>On May 14, 2001 at 12:57:07, Heiner Marxen wrote: >> >>>On May 14, 2001 at 08:47:07, leonid wrote: >>> >>>>Hi, Heiner! >>>> >>>>This time one personal demand. Can you send me your mate solver? I tried two >>>>times to do this by going into your site but something I did in a wrong way. >>>> >>>>I use Windows ME. If you have somewhere executive file, it will be the best. If >>>>whatever you will need to find in my program, you will just ask and my response >>>>will be instant. >>> >>>I myself do not offer any executables. Windows executables are not even >>>constructible for me. But Dann Corbit has made a WIN32 executable, >>>and offers all my sources, together with the EXE as >>> ftp://cap.connx.com/pub/chest/CHEST319.ZIP (674k) >>>If you want just the EXE (and the READMEs): >>> ftp://cap.connx.com/pub/chest/chest319.exe (268k) >>> ftp://cap.connx.com/pub/chest/README_LONG (30k) >>> ftp://cap.connx.com/pub/chest/README_QUICK >>> ftp://cap.connx.com/pub/chest/README >> >> >>Thanks, Heiner! Before I download not from this sources. I went directly to your >>site and can be that I download something for different system. >> >> >>>You should be able to download the above URLs with any browser or ftp program. >>>I suspect that Windows ME does happily execute WIN32 executables, so the >>>exe should be directly usable for you. >> >>Windows Me is not that simple to use. Or I myself screw somehow my Windows ME, >>or it is very prone to instability. It hung all the time. Last time just before >>I came here. > >Hi Leonid, > >I have no experience whatsoever with any Windows Version. > >Hi Dann, > >Can you tell whether the chest319.exe from your ftp site will run >on Windows ME without any trouble? Or whether it can be used from old DOS? >(Excuse me if these are stupid questions) I must admit that Windows ME is a botched bletcherous attempt at an operating system. Everyone I know who has installed it has despised it for many reasons soon thereafter. I have Windows 2000 on all of my personal machines and on my work machines. However, we do have WinME here and I have successfully run Chest on it. >>>Chest is a console application. For usage and examples please read README_LONG. >>> >>>>Now I am in changing few thing in my mate solver and for this I need some >>>>excellent mate solver as reference. It is not that simple to find as somebody >>>>could imagine. Before I had the impression that Genius 2 could do the work but >>>>later found that its moves number is fixed by 99 limit. The same is true for >>>>other best chess programs and even, to my susprise, for mate solver Mate 2.0. >>>>Mate 2.0 (he is in the package with Hiarcs program) do not permit to indicate >>>>directly number of moves that you would like for him to see. Its number of moves >>>>is once again limited to 99. He do not indicate the time when he think. >>>>It could be that I don't know how to use it. Hiarcs is not very talkative as >>>>programe in general. The best in its openess, between all program that I ever >>>>used, is Rebel. If I will one day end my chess program then I will try to make >>>>everything in my program as clear and accessible as this program did. >> >>>Is the limit 99 >>>(a) for the number of legal moves in a (single) position, or >>>(b) for the depth of the search? >> >>As far as I could understant it is limit for legal number of moves that initial >>position contain. As I could see, when initial position had 54 moves then this >>number was exactly 54. But when I asked for solving position that have initial >>moves 112 (just example), then number was indicated as 99. In the same time >>program indicate how many from those initial moves it already have seen. > >Fine, that is case (a). Chest is fine, here. > > >>>If (a): ok, Chest does better (222) >> >>I still must put my numbers as far. For now it is only 200. >> >>>If (b): sorry, Chest is limited even lower (30, with recompile: 63) >> >>If you speak about maximum number of moves to see, then you have one more >>advantange. Mine can go as far as 13. This is why all my positions never goes >>beyond 13 moves. Probably, I even will not change this limit very soon. > >Yes, that is what I spoke about. >And yes, 13 is nearly always enough. >That can change dramatically, once you have a hash table. >Then depthes beyond 20 are sometimes quite accessible, most notably >with pawn endgames. > >> My goal >>is to go to my chess part as soon, as I will find some way to speed my code when >>working with mate solver. Only when efficency of my chess part will reach mate >>solver code, then it will have sense to extend those 13 moves. For now my chess >>part is deadly slow. Its brute force can only slightly overrun other programs in >>10 plys search. In mate solver (forgetting about your that is wonderful >>exception) speed of mine for mate can go easely between 10 and 100 times beyond >>every best programs speed. I must absolutly recuperate part of it into my chess >>playing part. Mate solver is the best spot to see how much real speed you can >>obtain but not the first place to use it. I hope that you will do the same. You >>have bright chances to succeed. When you will do this, we will have even more >>things in commune to speak about. > >Do you want me to also extend my program to really play chess? >If so, I'm sorry. I have absolutely no plans to do so. I'm specialized >in mate problems, and as far as I can see, that will remain so. >May be some special ideas which I will implement for my mate solver are >adaptable to playing programs, but most probably I will not do so myself. I think it is a mistake to make chest play chess. I think what it does is entirely adequate and making a hybrid will be more work for little gain. Your planned extension to tablebase files will give me all that I should ever want from such a tool. I think (however) that Chest can be used to play chess, but not as some people might think. On a multiple CPU machine, you could have chest buzzing away on the root position. Often, it will salvage a mate that other programs simple won't discover for a very, very long time. Hence, it could offer advice to chess programs. This could make them play better chess. Also, general purpose chess programs could very much benefit from a mate solving mode. The methods of a program like chest could be employed in such a case.
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.