Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: New version of TSCP (1.8)--opening book and hash keys

Author: Uri Blass

Date: 04:42:16 01/30/03

Go up one level in this thread


On January 30, 2003 at 07:38:08, Janos Keinrath wrote:

>On January 30, 2003 at 04:57:31, Tom Kerrigan wrote:
>
>>A new version of TSCP is now available on my home page:
>>
>>http://home.attbi.com/~tckerrigan/
>>
>>A big change is the addition of an opening book. Many people have asked for this
>>feature but until recently I couldn't think of a way to do it that was
>>appropriately simple. The new version's opening book code is only 122 lines,
>>about half of which are comments. The book that's included with the program
>>(book.txt) has 385 opening lines. This feature should make TSCP much more fun
>>for humans to play against and it should also make TSCP stronger in the
>>tournaments that people run (I've heard of people "busting" TSCP).
>>
>>I also added hash keys to the program. I've said in the past that I would never
>>do this, but TSCP's old repetition detection code was pretty hard to understand
>>and I think it was a kinda buggy, too. So I was able to remove that and the new
>>brain-dead-simple hash code is < 100 lines. (It does slow the program down by
>>~10%, BTW.)
>
>I look at the source code and found it you always calculate the full hash key
>(with set_hash function) in makemove, but it not necessary.
>If you only 'xor' the from and to square info, and handle the special cases
>(capture,castle,promotion), it would be quicker.

There are a lot of ways to do tscp faster but the idea of it is to do something
simple and not something past.

Tscp also does not have a piece list that may make it faster and other functions
are also not optimized for speed.

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.