Author: Uri Blass
Date: 12:48:12 10/30/03
Go up one level in this thread
On October 30, 2003 at 14:20:55, Gian-Carlo Pascutto wrote: >On October 30, 2003 at 12:29:22, Uri Blass wrote: > >>On October 30, 2003 at 10:12:20, Gian-Carlo Pascutto wrote: >> >>>On October 30, 2003 at 10:01:50, Robert Hyatt wrote: >>> >>>>I didn't "pretend" do know anything. I simply said that such "slips" don't >>>>mean a thing. Tiger has been at the top of the SSDF. You haven't. Tiger >>>>has won more than one CC event. You haven't. That is pretty strong evidence, >>>>that is available in public, that suggests that such "leaked results" are not >>>>very reliable. >>> >>>With all respect for Christophe, but Chess Genius also topped the SSDF >>>list once. It doesn't mean a thing now, and the SSDF doesn't use duals either. >> >>Are we going to see Deep Sjeng in the ssdf? > >Working on it (it would help if the SSDF replied to my emails...) > >>The advantage from dual is less than the advantage from being twice >>faster(probably less than 50 elo), so the fact that the ssdf does not use a >dual does not mean that people can learn nothing from the ssdf results. > ><mode=Uri> >There is no proof that moving to a dual is less than 50 elo gain for Deep Sjeng, >so you cannot conclude anything from a SSDF result in regards to the relative >strength of Deep Sjeng vs ChessTiger on a single. ></mode=Uri> > >-- >GCP No proof but we can see the advantage from 450Mhz to 1200Mhz and 1200Mhz is usually more than 3 times faster based on tests because the advantage is not only the mhz and I remember that for some program like gandalf is was even more than 4 times faster . It is less than 100 elo for most programs so common sense suggest the advantage from dual at 120/40 is less than 50 elo(40 elo is my estimate). 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.