Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: about fruit2.2's bug(from an email that I sent Fabien in 9.10)

Author: chandler yergin

Date: 12:03:39 10/18/05

Go up one level in this thread


On October 18, 2005 at 14:59:51, Uri Blass wrote:

>On October 18, 2005 at 14:37:29, chandler yergin wrote:
>
>>On October 17, 2005 at 17:45:48, Uri Blass wrote:
>>
>>>On October 17, 2005 at 16:50:27, Majd Al-Ansari wrote:
>>>
>>>>Good job!  How on earth did you get that.  Observations such as this are very
>>>>valuable indeed.  I think the Fruit team should investigate this.
>>>
>>>simply I was curious about the ability of fruit to solve that study that I
>>>remember that many years ago no engine could solve.
>>>
>>>Fruit did not find that the wrong move is drawing so I was interested to see
>>>what is it's evaluation after the wrong move and I made it and noticed the bug.
>>>
>>>I decided after it never to use fruit for analysis unless the hash is clear
>>>because I do not trust results without clear hash based on that experience.
>>>
>>>If I make a move with fruit and want to analyze again I simply replace engines
>>>and replace it again to get fruit to analyze.
>>>
>>>Uri
>>        In the Analysis window Right Click on Engine name.
>>     You will get a Menu where you can clear Hash Table & Set Hash Table.
>
>There is this option but unfortunately the hash is not cleared after clicking
>clear hash tables.

 From Menu Click clear Hash Table first, then set hash, then select Engine.

>
>Fruit2.2 starts searching from depth 1 but get deep very fast.
>It seems that the only way to really clear the hash of fruit is to replace
>engines and replace again.
>
>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.