Author: Eugene Nalimov
Date: 11:29:21 06/03/99
Go up one level in this thread
Ok, let's be fair. Let's imagine the situation where orgnizers will give every participant quad Alpha 21264. Fair? After that some participants will refuse to play - "our program will not run there, because it's written in x86 assembly, or don't use more than one CPU, or don't tuned for such a speed, or...". My guess is that exactly the same people that say "it's unfair when the hardware is different" will say "Ok, but best commercial programs did not play, so winner is not World Champion". Or Fritz representative will say "Ok, I cannot run on your Alpha, so let me bring my x86 machine - I want to participate, even with disadvantage". What to do now? So, organizers decided to stay with x86, and will give each participants quad Xeon. Nevertheless, some participants will say "our program will not exploit four CPUs", so again there will be no fair play. And there will be chess program for a Sony Playstation, or Nintendo, of specialized chess board. And then there will be (crazy) person who wrote his program in PPC assembly for Mac. Why ban those program? Ok, let's ban them, and give single-CPU x86 to everybody. What now? Some will say "we spent a lot of time debugging parallel search instead of rewriting our program in x86 assembly, or we deliberately left our program in C, so it can be portable, and worked on other aspects of engine. Others wrote their program in x86 assembly. Now, why our hard work is not honored? We have a disadvantage". So, organizers decided to allow any micro, but not "big iron". After that somebody will say "they brought 32-CPUs Alpha, and it actually costs more than our System/3090 in minimal configuration. And their nps is higher. Why we are not allowed?". More, somebody will come with specialized chess chip, and he will say "it really costs less than even quad Alpha that is provided by the organizers". So, it looks that there will be unhappy people in any situation. With the current rules there is clear distinction - there is WCCC, WMCCC, and some time ago there was "equal hardware championship". Unfortunately, due to lack of financing (or maybe due to lack of qualified organizers, because old ones became tired, and next generation can only criticize) this year tournirs are "combined". IMHO, it's better to have not ideal solution than to have no solution at all. And if x86 commercial program will lose to some hardware monster, they at least can say "we have a hardware disadvantage", so in a sense everybody will be happy. Eugene On June 03, 1999 at 00:07:02, Prakash Das wrote: >On June 02, 1999 at 23:52:11, Prakash Das wrote: > >>On June 02, 1999 at 15:30:41, Tania Devora wrote: >> >>> >>> >>>What is the hardware for the 9th Wccc99 ? >>> >>>There are many strong computer programs that not go to participate in this >>>World championship , like Chessmaster6000, and many others. >>> >>>I think that this is not a fair play . Cilkchess and others programs will run is >>>a Super machines, is evident that could win easy. >>> >>>For me all the programs, will run in the same computer with the same ram memory. >>> >>>Is the only way to see the real strengh of the program. >>> >>>For example Fritz5.32 in a Pentium III 500 MHZ could beat Fritz5.32 (pentium 200 >>>mmx) , for a big score. >> >> >> Tania, you are wasting your breath. A while back, I started a thread called >>"Uneven hardware for wmcc?" in which I questioned the purpose of this >>self-congratulatory exercise. >> Most of teh replies went like this: it's fun, it's like creating a Mount >>Olympus of chess programs, we are coming with the meanest baddest machine, etc >>etc. No one really cared about addressing the real point: what the hell is going >>to be accomplished. >> So, sit back, and get ready for a few yawns. You will see the same old programs >>which are at "top of charts" decimate the others (not bad programs necessarily) >>running on weak hardware. >> If we tried to solve cancer in this "scientific" manner, there's no hope for >>living beings. > > Well, following up on myself. I thought this was obvious, but as in my previous >post on that other thread, let's say program A on big bad hardware B, beat >program D on weaker hardware E. What did this prove? Wins and losses are all >relative, and so will be the standings later on. What kind of Mount Olympus did >we scale (Karinsdad)?? It is a mirage where we appear to have created a super >"playing system". But this would have achieved by beating another on weaker >hardware. Next day, the result will be different. > > Let's hold a WMCC everyday with changing hardware. That way, we will create >tons of mount Olympus!
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.