Author: Chessfun
Date: 11:32:57 10/18/03
Go up one level in this thread
On October 15, 2003 at 15:28:11, Peter Skinner wrote: >On October 15, 2003 at 13:25:59, Aaron Tay wrote: > >>On October 15, 2003 at 11:11:34, Peter Skinner wrote: >> >>>On October 15, 2003 at 10:23:56, Thorsten Czub wrote: >>> >>>>it is as old as computerchess, we were able to measure it in the old times >>>>of dedicated chess computers, and it has IMO not changed on the PC: >>>> >>>>chess programs do profit from time DIFFERENT. >>>> >>>>some are better on fast levels. others are better on slow time controls. >>>> >>>>As far as i remember there are only a few people outside believing that >>>>the programs strength is linear increasing function concerning time controls / >>>>cpu-speed. >>> >>>In Chess Tiger's case I have matched it against itself using two different sets >>>of hardware to see how well it scaled. >>> >>>I used a PIII 900 vs a PIV 2.6 Ghz. The match of 200 blitz games ended in a dead >>>draw. The reason I did the test was that I have used Tiger since version 11.9, >>>and my hardware gets upgraded atleast once a year. I have noticed no rating >>>difference while playing online when using different hardware. >>> >>>Other programs like Crafty, Ruffian, Gandalf do indeed profit more from better >>>hardware. I just don't think that Tiger is one of them. This could possibly be >>>the reason for Christophe's thoughts on the time control difference. >> >>>Peter. >> >>The last doesn't make sense, if Tiger does not gain from time control >>differences, but it's opponents do, then Tiger's relative performace would vary, >>and Christophe would indeed notice :) > >What I meant was when I upgrade my hardware, and everyone else's stays the same >(for a small time atleast) I do not see an increase of strength. > >Take two computers with different MHz, search for the same position and the no >matter what.. each computer comes up with the same. The nodes may vary, but the >time it took to search did not. That is of course totally wrong. The nodes to the solution would be the same or very close and the solution time would be totally different. >I have done this many times with Tiger. There is >no real strength gain when using bigger hardware, or none that can be very >noticable. Then please supply me a position and I'll prove it to you on an XP 2600 and a TB 1200. Also the results by the SSDF themselves prove it look for engine using 450 and 1200 mhz and see the difference yourself. Sarah.
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.