Author: Robert Hyatt
Date: 14:38:16 07/21/00
Go up one level in this thread
On July 20, 2000 at 19:57:14, Ralf Elvsén wrote: >On July 20, 2000 at 18:36:25, Chris Carson wrote: > >> >>1. DT scored 90% against micros from 1988 to 1995. DB is >> at least 100 times faster than DT and todays fastest >> micro's have not closed the gap (P6-200 was released >> in Nov 1997). >> > >I don't want to enter your discussion, just ask a question. >What was the hardware of DT during this period (88-95)? In particular, >how did it evolve? > >Ralf It started out in 1987 as a single chip chess machine searching 300K nodes per second. I believe that by 1988 it was searching 700K nodes per second with new hardware features. 1989 saw it run with two processors (I think). They slowly added more hardware processors, up to a max of 20 I think, but they rarely used all 20 in matches due to flakiness and heat sensitivity among other things. During this time the search was modified heavily and the term "Singular Extension" was born (an idea of Hsu's). They did lots of neat things including finding a better way of recognizing when they were in trouble, rather than just relying on a "fail low". And no, I don't know how they did it as we never had time to sit down for a long while to discuss it. It improved each and every year, although I don't think they re-did the chess processors more than one time, until they did the original DB chip and then the DB2 redesign of the chip.
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.