Author: Robert Hyatt
Date: 14:17:27 09/19/99
Go up one level in this thread
On September 19, 1999 at 15:04:53, Christophe Theron wrote: >On September 19, 1999 at 12:13:03, Will Singleton wrote: > >>On September 19, 1999 at 10:12:59, Robert Hyatt wrote: >> >>>On September 18, 1999 at 22:49:20, Christophe Theron wrote: >>> >>>> >>>> >>>>I think a tournament on slow hardware (386-16 to 386-40, or 486 <=33MHz) >>>>including top programs of the early 90s and programs of today would be very >>>>interesting. We could amongst other things see if there have been really no >>>>advances in software in the last 10 years. >>> >> >>>I don't think the test would be 'fair' if that is what you want to examine. >>>IE many of the software improvements come as a result of hardware changes that >>>make the programs fast enough that they can do things they couldn't on slower >>>hardware. IE would you play _anyone_ if you could only do a 5 ply search? >>>Would you even think about null-move R=2? Would you have your program spend >>>50% of its time in the evaluation? THose are all decisions I had to address >>>and the answer would be different if I was doing 1K nodes per second on a >>>386/16.. >>> >> >>My program (since it only plays on one machine) definitely is targeted to that >>speed (300 mhz). It wouldn't be the same program if it had to play on a 68000 >>25 mhz. >> >>However, I'd like to know if some of the commercial programmers make provisions >>for different targets, and alter search and eval strategies based on machine >>speed. >> >>Will > >I think it's a very bad idea to target for a given speed. Sorry, but I see no >reason to do this. When I make a change in Tiger I make various tests on >different computers to be sure that nothing is broken. These tests include blitz >games against Genius5 on my 386sx20 and longer games on a 300MHz computer. > >I think that making sure that the strength is unaffected at very different >speeds is a way to make sure that you are going into the right direction. At >least I think it works for me... > > > Christophe I personally believe that doing this is impossible. IE I can't imagine a program that plays equally well if it searches to 6 plies or to 10 plies. I had this problem for years as my Cray Blitz development was mainly done on a VAX, and then we would run on the Cray for tournaments. And on occasion, it was very obvious that things that were helping at 6 plies were killing us at 10... just my opinion, of course. But you certainly can't get away with null move R=2 on a 386. It will be so blind to king-side attacks that a good expert will eat it alive. I saw problems on a P5/133 with Crafty. They don't show up near as often on today's hardware...
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.