Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Mate in 1 - but Fritz 6 needs 1 hour!!!

Author: Heiner Marxen

Date: 10:00:35 08/14/00

Go up one level in this thread


On August 14, 2000 at 04:20:18, Colin Frayn wrote:

>On August 13, 2000 at 14:54:29, Heiner Marxen wrote:
>
>>>Have you seen how I do hash tables?
>>>
>>>*grin*
>>
>>I'm not sure, what the *grin* indicates, agreement or disagreement.
>>Could you clarify?  Do you hash exact positions?  (unlike everyone else,
>>except perhaps me?)
>
>Yep.  Well, I use two separate hash keys, both with an extremely low collision
>rate, making the combination extremely reliable.

I see.  Now, there is still no guarantee for not erring.
I understand that you neglect the possibility of error, here.
With full 64 bits of a good hash this appears to be quite reasonable.
OTOH, every "mate-in-xy" statement from the program is to be taken
with a (small) grain of salt... but that is true for most (all?)
of the other programs.

>>Now, that starts to get quite interesting for the author of a mate solver.
>>Would you care to share some details?
>
>Heh ;)  Well I read your next message so have a browse through the code and see
>if there's anything of use in there.
>
>Feel free to email me if you have any questions/suggestions.

Sure, thanks.  Currently I am very busy working on a non-chess topic
(busy beavers), so it will take some time before I come back to work
on Chest (sorry Dann) or other chess related stuff.
But I promise I will have a closer look to your mate (and check) detection
code.  May be I can make some suggestions.

You said your mate detection code is only a small fraction of the overall
time, so speeding up this code will not have any noticable effect, anyhow.
(Provided it really is only a small fraction ;-)

>Cheers,
>Col

Cheers until later,
Heiner



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.