Author: Christophe Theron
Date: 06:18:28 01/07/02
Go up one level in this thread
On January 07, 2002 at 05:56:55, Bas Hamstra wrote:
>>I think there can be many factors for chess contests. You can have open
>>hardware, or fixed hardware. On open hardware, you can equalize the time.
>>
>>It depends on what you are trying to determine. For instance, do you want to
>>know what the strongest combination of hardware + software is?
>
>In fact when I think about it, I can't see the point of this. IMO if you would
>ask all programmers what they think is more important a) the competition aspect
>or b) determining what is the strongest "chess entity" 99% would vote for a.
>
>>OK, now let's think of equal platform.
>>
>>Joe writes a program for the Mac.
>>Sam writes a program for Linux on Mips
>>Sally writes a program for OS/2
>>Jill writes a program for Alpha on NT
>>Larry writes a program for Win32 on Intel
>>Fred writes a program specially optimized to use AMD's registers
>>
>>"Just say everyone must use Windows" simply disallows the Mac programs (which
>>are popular) and even the Linux programs. Similarly for everyone must use the
>>same hardware. So that any decision you make necessarily will hurt *someone*
>>in whether they can perform or not, or at least infringe upon their ability to
>>perform.
>
>I have seen these arguments, but they absolutely don't convince me. This kind of
>hurting is *nothing* compared to the hurting that is currently going on, where
>someone on a P3-700 has to compete with a fast quad machine in a tournament.
>
>>I think (however) that equal hardware events can be very interesting. You
>>could use the results of the SPEC benchmarks to normalize for any tested
>>machines, and then give time slices which allow for the differences.
>>I don't know of any communication protocol which allows that sort of time
>>control yet, but it should not be impossible to create it. You can (of course)
>>manually step through a move at a time (like the KKUP and KKUP2 contests) but
>>that would be far too tedious for normal game speeds.
>
>Possible, but complicated. A good compromise is IMO what they do in Paderborn:
>the organization aranges all hardware, and you can for instance chose P3-700
>single or P3-700 double for your program. Nothing is perfect, but pretty fair I
>would say!
>
>For tournaments in which we have to bring our own hardware we are in a vicious
>circle. Since everybody is bringing the fastest hardware on the market, I have
>to too. Why not end that, and set a speed limit. Everybody can arrange an old
>laptop, very convenient, and we could compete in fairness forever.
>
>Best regards,
>Bas.
I support this idea totally, but I know that other programmers are afraid of
competing with equal weapons.
Personally I would be very glad to enter an uniform platform tournament. But
this is something that we talk about since a long time and it never happens (at
least no World title has ever been played under these conditions, and I'm pretty
sure it will never happen).
Christophe
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.