Author: Robert Hyatt
Date: 19:17:41 09/11/99
Go up one level in this thread
On September 11, 1999 at 17:20:44, odell hall wrote: >On September 11, 1999 at 16:23:30, Pete Galati wrote: > >>On September 11, 1999 at 16:07:47, odell hall wrote: >> >>>On September 11, 1999 at 15:27:43, Robert Hyatt wrote: >>> >>>>On September 11, 1999 at 14:05:12, Leon Stancliff wrote: >>>> >>>>>Robert, >>>>> >>>>> >>>>> >>>>> I have been speculating for a long time on what would take place in a match >>>>>between the top ten computer programs running on easily available machines and >>>>>the top ten chess players in the USA. >>>>> >>>>> I have been playing my Hiarcs 7 program on a Macintosh G3 at 275 Mhz. It plays >>>>>almost even with both Data and Singacrafty on ICC. >>>>> >>>>> I have assembled the following list of computer programs from such sources as >>>>>the SSDF list and the Selective Search list. In your opinion, would Data, >>>>>Singacrafty or Crafty deserve a position in this list? >>>>> >>>>> >>>>> >>>>> Hiarcs 7.32 >>>>> >>>>> Fritz 5.32 >>>>> >>>>> Chessmaster 6000 >>>>> >>>>> Nimzo 99 >>>>> >>>>> Chess Tiger 11.8.2 >>>>> >>>>> Junior 5 >>>>> >>>>> Rebel 10 >>>>> >>>>> Shredder 3 >>>>> >>>>> Genius 5 >>>>> >>>>> MChess Pro 7 >>>>> >>>>> >>>>> >>>>> Let's assume that each is running at about 550 Mhz on a single CPU. >>>>> >>>>> >>>>> >>>>> MChess Pro 7 >>>> >>>> >>>>I wouldn't venture a guess. If you put crafty on a quad xeon 550, it can beat >>>>_any_ of those programs. On equal hardware I really don't know as I don't play >>>>on 'equal' hardware. :) >>> >>> >>> Oh yes!, i would love to see the results of crafty verses the top programs on >>>equal hardware, It would be interesting to see how it would measure up. >> >>Consider that putting Crafty on equal hardware takes away one of it's valuable >>tools, probably not fair to the nature of Crafty. >> >>Pete > > Well, I would think that if the purpose of pitting one program against another >is to see which program is written best, and has the best playing strength, then >equal hardware is required, otherwise the test is not valid You miss the point. Suppose we decide to play a match of Crafty vs a currently available commercial engine. And we make the equal hardware an 8-way xeon 550 box? Not fair because crafty can use all 8 and none of the commercial programs can? Why? Suppose we choose the alpha 21264 as the equal hardware, since that is the type of architecture Crafty is designed around. Not fair? IE there is _no_ way to do this reasonable. If you use one cpu, you toss out a year's work I have done. If you use 4 cpus, you give me a big time advantage. Either way is "wrong" in the concept of just testing the software. Because my parallel search _is_ part of the program.
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.