Author: Robert Hyatt
Date: 14:22:03 05/04/01
Go up one level in this thread
On May 04, 2001 at 14:46:41, Sven Reichard wrote: >On May 04, 2001 at 09:25:21, Robert Hyatt wrote: > >> >>I don't know where that comes from but DB was _not_ a "root processor" by >>any definition of the word. The evaluation was 100% tip evaluation... that >>was the purpose of the hardware in the first place.. > >I don't know if they changed it for the later machines, but Deep Thought fits >the description of a "root processor" in the following way: > >The tip evaluation was basically based on piece/placement tables (actually, it >is a bit more involved, in the sense that there were more tables than just one >for each piece. All in all, A. Nowatzyk speaks of "nearly 40,000 programmable >parameters"). These tables were loaded before each search based on (and that is >the important thing here) features of the root position. This explains >difficulties in evaluating transitions as described by Uri. > >Regards, >Sven. My point is that I don't think they were root processors in any current sense of the word. The ability to change parameters and reload them was important to the design. But I don't believe they did this after each move. Although I could see somethings that might change (ie you don't have a knight on the board so some things could be modified to reflect that). But I don't think simple material on the board was used to change things from move to move. If so, you would have to totally clear the hash table after each search, and they didn't do that that I ever saw...
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.