Author: Dann Corbit
Date: 18:12:45 06/17/02
Go up one level in this thread
On June 17, 2002 at 20:28:30, Robert Henry Durrett wrote: >On June 17, 2002 at 20:18:31, Dann Corbit wrote: >>On June 17, 2002 at 20:13:26, Robert Henry Durrett wrote: >>>I recently contacted Chessbase [info@chessbase.com] and received an email >>>response saying that they will soon provide the information I requested, and >>>more, about the computer. It will appear on their Chessbase News webpage. >>> >>>In the meantime, it's not too late to speculate! >>> >>>While discussing the computer [for the Kramnik match] with my son, who is here >>>at my house for a visit, he noted that "there are plenty of 64-bit 64-processor >>>computers running under UNIX." When pressed for details, he could not give me >>>any because he does not use such "beasts" in his job. Maybe someone here is >>>more familiar with this? >>> >>>The commercial version of Fritz runs on PCs that have Windows operating systems >>>simply because the customers want to put Fritz on their home PCs, which run >>>under Windows. >>> >>>But is there any reason, or LAW, that says that the computer they will use for >>>the Kramnik match cannot run under UNIX? >>> >>>Essentially the same situation applies for word length. The Fritz customers >>>have 32-bit computers, so Fritz is designed to run on those 32-bit computers. >>> >>>But is there any reason, or LAW, that says that the computer they will use for >>>the Kramnik match cannot be a 64-bit computer? i.e. have a 64-bit deep fritz >>>running under UNIX? >>> >>>If so, do any of the "UNIX-types" here at ICD/CCC know how fast the processors >>>could be in that case? Which processors would they use? >>> >>>What do you think? Is an 8-processor 64-bit computer running under UNIX a >>>possibility for the Deep Fritz versus Kramnik October 2002 match? >> >>Not a snowflake's chance in the proverbial place of eternal punishment. > >Do you mean "Hell"? > >> >>The whole purpose of the match is to sell computer chess programs. The >>ChessBase company wants to recoup some of the zillion dollars they are putting >>on the line, and I certainly do not blame them. This is not a pure scientific >>experiment with a desire to find out facts. It is a way to make money. > >My gut feel [intuition] makes me feel that you are right, even though I really >don't want it to be so. I don't know why not. If money were not involved, the experiment would not take place. And nothing would drive forward towards excellence except the thirst for knowledge. But the thirst for knowledge will not fund a ten million dollar project. >>Now, nobody is going to buy 64 bit Fritz for Linux. OK, maybe one in a million >>of the sales they would have for Win32. Can anyone seriously imagine that they >>would run it on anything except a Win32 platform? > >But I just did! But was it serious? >>Now, they might get an 8 CPU 64 bit version of Windows and use that. Most Win32 >>users will have no clue that that system is a trillion miles from theirs and >>they can still use the performance to drive sales. > >True. Most computer users are really not up to speed on the guts of computers >and software. > >There is a counter argument: The Fritz guys may REALLY WANT to win this match! > >If they could get a much more powerful computer by using UNIX [or LINIX???] than >they could get by using windows, then they might be tempted. > >Since you brought up the subject of Hell, I am reminded of stories about people >making deals with the "head man" in Hell. [I am reluctant to mention his name.] > >Maybe such a deal? :) > >Seriously, is it not true that the switch from Windows to Unix would allow them >to use a much faster and more powerful computer? [Faster processors, etc.]. >The schedules published by Intel and others do not seem to allow for really fast >8-processor computers running under Windows in the near term. At least, that is >my understanding. For SMP machines, the 64 bit unix machines using a sledgehammer or mckinley chip can just as easily be made to use Win32 as UNIX. There are multiple CPU versions of Win32 and versions for 64 bit chips already made. There are 32 way Alpha boxes, but I think they cost too much for this experiment and would not drive sales much for ChessBase so the incentive to use one is zero. Actually, I would be pretty stunned if they did choose a route like that.
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.