Author: Danniel Corbit
Date: 17:47:05 11/05/97
Go up one level in this thread
Notion to handycap computers... Put the same game on two different computers and have it play itself 10 times. Do the same thing with all the other programs that run on those two architectures. Form a least squares fit of the machine's abilities. Do this for each different system, with one as a baseline. Even with all of that, I think that such an analysis would have limited value. Perhaps machine "A" has a large cache that changes its ability under different circumstances of play. Perhaps program "Y" is tailored to the assembly language for machine "Z" and performs better on that system than its raw performance would dictate. It might be amusing to pit the winning program against itself on several different architectures, if such a thing were possible [might be problematic for Mac only programs or the like]. But to see Junior/P300 face Junior/Alpha767 and the winner take on the winner of Junior/K266 vs Junior/PowerPC might be equal fun to the original tournament. This contest clashes both different programs and different machines at the same time and I think that makes it even more exciting. Look at the buzz over the Alpha machines. I am quite surprised about how very well the P300 chips did. I think an even baseline contest is another interesting idea, but that will lock out people who do not target the baseline platform.
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.