Author: Uri Blass
Date: 08:48:40 07/14/04
Go up one level in this thread
On July 14, 2004 at 11:11:44, Sean Empey wrote: >On July 14, 2004 at 02:15:31, Uri Blass wrote: > >>On July 13, 2004 at 15:00:36, Sean Empey wrote: >> >>>On July 13, 2004 at 14:54:17, Uri Blass wrote: >>> >>>>On July 13, 2004 at 14:32:14, Sean Empey wrote: >>>> >>>>>On July 13, 2004 at 14:16:54, Uri Blass wrote: >>>>> >>>>>>On July 13, 2004 at 14:03:14, Sean Empey wrote: >>>>>> >>>>>>>On July 13, 2004 at 13:40:37, steven blincoe wrote: >>>>>>> >>>>>>>>>Hardware is part of the program. >>>>>>>>How would you test the strength of dedicated units when their hardware veries so >>>>>>>>much? >>>>>>>> >>>>>>>>its true to some extent..sometimes i test computers with different hardware but >>>>>>>>the resluts will generally always be true to the hardware >>>>>>>> >>>>>>>>the stronger the hardware used in a dedicated unit ,then the more powerful the >>>>>>>>rating >>>>>>>> >>>>>>>>there will be no computer running on a Motoroloa 68000 32 bit processor that >>>>>>>>could lose to a computer running on an 1806 processor >>>>>>>>no matter how powerful the code for the chess program >>>>>>>> >>>>>>>> >>>>>>>>but every year it seems we have the same resulting discussions for the WCCC... >>>>>>>>this program lost because it was not running on a dual operton..or ..if we >>>>>>>>used a qaud opteron we would have preformed better..etc.etc. >>>>>>>> >>>>>>>>Steve >>>>>>> >>>>>>> >>>>>>>Yes, the discussions happen but they are IMHO, useless questions. Asking if >>>>>>>Program A would run better on a quad is not reasonable if it is not a MP >>>>>>>program, etc. I think people like to speculate how strong a certain program >>>>>>>_might_ be, but their is a huge difference in MP implementations and pulling it >>>>>>>off. If a program is not 64-bit then asking if it would have done better running >>>>>>>on 64-bit architecture is also useless. Just my two cents. >>>>>>> >>>>>>>-Sean >>>>>> >>>>>>The problem is not only hardware. >>>>>>Some programs were compiled by inferior compilers. >>>>>> >>>>>>Movei was compiled by Visual C++6 except the first round when I used a version >>>>>>compiled by Dann Corbit but I found that there was a bug in time management with >>>>>>ponder on and I also wanted to change other things during the tournament so I >>>>>>used again a version that is compiled by Visual C++6 in the rest of the games. >>>>>> >>>>>>I was surprised to hear that Woodpusher was even compiled only by Visual C++5 >>>>>>The programmer did not buy better compiler because it was too expensive for him. >>>>>> >>>>>>I do not think that it is a fair advantage. >>>>>> >>>>>>Uri >>>>> >>>>> >>>>>So now people should be forced to compile in a specified compiler? It seems to >>>>>me people are trying to cripple other's programs so that those programs that >>>>>aren't as strong still have a chance to win. How is that fair? I don't see how >>>>>handicapping everything is fair. >>>> >>>>I do not think that handicapping other is fair. >>>>I think that the most fair situation is that the sponsor also gives the >>>>participants the best compiler to compile their program. >>>> >>>>Teaching a program to use a parallel search is a lot of programming effort but >>>>compiling with a better compiler should not be a lot of effort >>>> >>>>Uri >>> >>> >>>I'm sorry, but this makes no sense to me. I don't expect anyone to give me a >>>free compiler. What about those on *NIX, Windows, MAC. If the new MS compiler is >>>faster, then it will be said that the reason the program won is because of the >>>compiler? >>> >>>Man, there's more conspiracy theories, What if's..., accusations, "this isn't >>>fair", "but he has...", blah, blah, blah in CC than any other close activity. >>>How about folks just take their losses and wins with some dignity. >>> >>>-Sean >> >>Saying that something is not fair is not blaming a specific person. >> >>I think that every advantage that is not about programming is not fair if people > >Last I checked a threading program is about programming. Even if you all ran on >equal hardware, you probably would still have lost. Face it, you were out >classed. Complaining doesn't fix that. Spend the energy improving your program >for next year. movei isn't a slacker by any means. Even if I was to join WCCC; I >wouldn't because the amount of whining is on par with a 3rd grade science fair. > >-Sean I did not claim that parallel search is not about programming. I did not claim that movei did not win because of hardware. Movei is still weaker than Crafty even if crafty is handicapped by running on only one processor and Crafty was not the best program in the tournament. I claim only that the situation is not fair if people look at it as a competition of programming because programs that could use better hardware like Deep Sjeng or ParSos used only a single processor not because of programming advantage and programs that used a single processor did not use the best one not because of programming. Uri
This page took 0.01 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.