Author: Georg v. Zimmermann
Date: 02:44:45 08/11/03
Go up one level in this thread
On August 11, 2003 at 00:24:07, Federico Corigliano wrote:
>Hi to all
>
>I want to implement the option info hashfull <x> of the UCI protocol, and for
>that I need a way to calculate the hash saturation.
>Looking for some Internet pages, it can be calculated (I think) doing:
>used hash entries/total number of hash entries. With a single ("replace if same
>depth or better") hash table the number of used hash entries can be calculated
>with a counter that is incremented each time that I store info in a hash entries
>that don't have previous info.
>But I have a problem with this. I have a two-level hash table and in one of them
>(the "replace if same depth or better" table) I have an "age" counter, so I can
>replace hash entries with info of an ancient search. My previous trick to
>calculate the number of used hash entries will don't work well here.
>How I can do it?
>
>Greetings,
>Federico
I'd count overwriting entries from previous searches like storing into empty
positions. So regarding the hash statistics, the hash table would be "empty" at
the start of every search.
Georg
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.