Author: Dann Corbit
Date: 10:49:37 06/24/04
Go up one level in this thread
On June 24, 2004 at 07:55:48, Paul Clarke wrote: >On June 23, 2004 at 23:06:13, Dann Corbit wrote: > >>On June 23, 2004 at 22:50:38, Mike Byrne wrote: >> >>>Do we have now have machines on desktops equal to Deep Blue in strength as Don >>>Daily suggested may happen in 10 years or less back in 1999? >[snip] >>No question about it. >>At the half-way point, an 8-way AMD 64 bit box gets a significant chunk of it >>(about 10% of Deep Blue's best burst rate IIRC). >> >>with 5 more doublings, that would give 1.6x Deep Blue's full tilt, max >>throughput. Of course that is for bleeding edge. A high-end workstation will >>be about on par with Deep Blue, I would think. > >Five more doublings would be 3.2x Deep Blue, surely? 10% ==> 20% 20% ==> 40% 40% ==> 80% 80% ==> 160% 160% ==> 320% Quite right. That's what happens when I do it in my head. >Though expecting five >doublings in five years seems optimistic: the doubling period in Moore's Law is >usually given as 18 months to two years, so three doublings seems more likely, >giving 80% of Deep Blue's peak performance. It is super-exponential. The acceleration is also accelerating. It now doubles once per year. > I've also seen articles suggesting >that Moore's Law might finally be running out of steam, as chip designers have >started to hit problems with reducing feature size. Those articles are wrong. Look at this one: http://www.kurzweilai.net/articles/art0408.html?printable=1 And similar stuff that you will find here: http://www.kurzweilai.net/index.html?flash=1
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.