Author: Robert Hyatt
Date: 06:38:06 08/28/00
Go up one level in this thread
On August 28, 2000 at 09:12:08, Wayne Lowrance wrote: >On August 27, 2000 at 10:23:40, Robert Hyatt wrote: > >>On August 27, 2000 at 02:01:20, Christophe Theron wrote: >> >>>On August 26, 2000 at 23:58:38, Robert Hyatt wrote: >>> >>>>On August 26, 2000 at 21:14:00, Christophe Theron wrote: >>>> >>> >>>(snip) >>> >>>>>>Restricting things to one piece of hardware is stifling, not leveling. >>>>> >>>>> >>>>>I don't see why competing with equal weapons, especially when it is the platform >>>>>99% of people on earth is using, is such a problem. >>>> >>>>Because you are wrong. 99% of the people on earth don't have the same >>>>processor at the same clock frequency with the same amount of memory and >>>>disk space. Some still have P6/200's (I have one at home and a lab full >>>>at the office). Some still use 486's. A few have 1ghz PIII's. So the >>>>world is _not_ uniform platform. If you stick on X86, do you let 600 and >>>>1.3 ghz processors in? Do you make everybody use 1ghz processors? Is >>>>that what everybody in the world is using? >>>> >>>>I think this line of reasoning is simply flawed, in a very basic way. >>> >>> >>> >>>Yes and no. >>> >>>Follow my reasonning: >>> >>>99% of people are using the x86 architecture, so let's make it a x86 platform >>>event. So it means something to most people. >> >> >>I have a P6/200 at home. Explain to me how it is any different for you to >>play on a PIII/1ghz which will cost around 3,000 dollars, and my playing on >>a digital alpha, which will cost about 4,000 dollars? Both have no relationship >>to my computer at home. Both cost a significant amount of money. As far as the >>price difference between them, just compare a gateway/Dell price to an IBM price >>and you will get the same difference. >> >> >> >>> >>>Then if we make a x86 platform event, so why don't we make it a uniform speed >>>uniform platform event? If somebody agrees to run his program on the same >>>platform as the other ones, why wouldn't he agree to run at the same speed? >>>Unless he knows that his program is weaker, of course. >> >>Or unless he knows that his program was designed for a particular architecture, >>and running it on the PC would not provide optimal performance. ChessMachine >>is one obvious example. Old versions of programs written in assembly are >>another one. Who in their right mind would prefer to program a X86 in >>assembly when they had the choice of (say) a M680x0, or a sparc, or an >>alpha? >> >>So either everyone has to start with the X86 as the target platform, or they >>get excluded from such events... >> >>> >>>One thing leads to another. Once we agree to run on a given platform, it seems >>>natural that everybody runs at the same speed. >>> >>>Because what most people are interested in is "what program is the best", not >>>"who is able to put the bucks and come with the fastest hardware". >> >> >>This is not true. Just check with the operators of chess engines on ICC. IE >>people like Lonnie bought kryo-type machines, just to have the fastest box they >>could get their hands on. There is always a group of people looking for max >>performance. But this still doesn't address the hideously stifling approach >>that using a uniform platform would address. IE who would have ever tried >>bitmaps if 32 bit machines were all we would ever have? >> >> >> >>> >>> >>> >>> >>>>>Maybe you forget that there are people out there using our chess programs. And >>>>>these people are using x86 computers, mostly. >>>> >>>>So? In the USA, most people drive a GM product. Do we not allow Chrysler and >>>>Ford into the auto races based on that reason? >>> >>> >>> >>>Your comparisons with cars and sports are getting boring. >>> >>>Come on, I know you can do better. I could have predicted your move... >>> >>>Can I help you? Why don't you take for example the evolution of the animal >>>species? >>> >>>Obviously they have not chosen to be "uniform platform". Aha! So uniform >>>platform is stupid. Period. >> >> >>I don't see any uniform platform. I see warm-blooded animals, cold-blooded >>animals, animals that eat meat, animals that only eat plants, animals that walk >>on two feet, animals that walk on four feet, animals that can run 80 miles per >>hour, animals that can barely run 4 miles per hour. You _really_ think that >>the animal kingom is "single platform". They all are based on carbon, all have >>some form of blood. Hey, computers are based on silicon. All use electrons. >>Is that a trend? Based on your comparison with animals, all computers are a >>uniform-platform. >> >> >> >>> >>>But anyway, all these comparisons miss the point. >>> >>> >>>My point is that most people are interested to know how programs perform on the >>>kind of computers they have. >> >>OK... but we aren't showing them that. We aren't showing the folks with >>P200/mmx's how the programs would compare there. We won't be showing the >>_majority_ of people how the programs would compare on their specific type >>and speed of processor, only how the programs do on one particular architecture >>(X86) at some specific speed that is probably way faster than what they have at >>home. > >I have been tempted to comment several times on this thread. I am a program >purchaser. I have many of top current programms. Bob I find my self in general >not in favour of your thoughts on uniform platform. The above paragraph in >particuliar is flawed. When we buy a program It is in _our interest_ that all >are playing on the same platform. The fact that the uniform platform would most >likely be running at higher processor speed is ok, _it is uniform_ So if all I >can afford is a ole P200 that, is my restriction. But I am left to guessing if I >have to make a choice between one program running on a super cooled widget or a >8 way Xeon etc verses others running on inferior hardware. I don't see the difference. You have a P200/mmx. I can run on either a 1.4ghz PIV, or a 800mhz alpha. Why does it matter in your decision-making? Will you actually buy a PIV/1400mhz machine? You could buy an 800mhz alhpa as well. I don't have anything against uniform platform at all. But for the "main tournament" between computers, choosing uniform platform is a stifling move. Look at _all_ the developments over the years. How many came from uniform platform programs and how many came from heterogeneous architectures? We have had micro events since 1980 that included X86, 6502, 68000, Z80, ARM, and more recently even the PPC (macintosh). Is there _really_ a good reason to exclude the PPC/macintosh? If so, what? 20 years ago should we have excluded all the mephistos, sargons, chessmachines, etc??? On the contrary. At most computer chess events, the _largest_ group is the amateurs. They aren't selling anything. > >Most of these programs that are entered are _marketed for sales_. That means it >is important to consider what is best for us. > >Your non uniform platform has it's merrit. I am very interested in the best >performing Computer chess against human GM's. Here lets take our best shot as in >the DB vs KC. > >Regards >Wayne >> >> >> >>> >>>Let's take a PC of reasonnable power, give it to everybody, and see what >>>happens. >>> >>>The idea is simple. I can understand it is not to your taste, but would >>>certainly help to get more attention from the public audience. >> >> >>I don't see why public interest would increase. Again, you would think that >>if this was something that everybody wanted to see, the ACM events would have >>evolved into uniform platform. Don Beal used to run one, but quit doing so. >>For lack of interest, I assume. >> >> >> >> >> >>> >>> >>> >>> 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.