Author: Robert Hyatt
Date: 07:02:26 08/10/01
Go up one level in this thread
On August 10, 2001 at 04:15:15, Graham Laight wrote: >On August 09, 2001 at 22:09:29, Robert Hyatt wrote: > >>On August 09, 2001 at 20:41:30, Vincent Diepeveen wrote: >> >>>On August 08, 2001 at 12:50:07, Graham Laight wrote: >>> >>>>I don't think that xx86 --> Pentium range makes a good choice for computers with >>>>(very) large numbers of processors. IMO, large numbers of processors represents >>>>the best future option for increasing computer power in a cost-effective way. >>> >>>Have to disagree directly. No one can afford so many processors, >>>so much transport costs, so many risks, such a big power bill. >>> >>>Easier is a single processor being way faster and pressed cheaply > >This line of argument is ridiculous. For a start, it precludes the possibility >of putting multiple processors on a single chip. > >Even if multiple processors have to be on multiple chips, it's not going to >break the axles of transport trucks, result in death by exposure to >electomagnetic radiation, or drain the power grids (except in California :) ). > >>The math doesn't work. For any single processor you build, I will take >>two of 'em and make a dual for less than 2x the cost. And I will be twice >>as fast for that <2x cost. Ask intel what it will cost you today if you want >>them to build you a 4ghz processor. Hint: You won't get it for 2x the cost >>of a 2ghz processor. You won't get it for 10X. You won't even get it for >>100X. >> >>SMP is here to stay, from an economics point of view. >> >> >> >> >>> >>>>How a good multiprocessor chip can become a new industry standard right now is >>>>difficult to see. >>>> >>>>One way it could possibly happen is if ordinary people stopped buying power >>>>machines and went for cheap ones (a sensible choice IMO) - but server purchasers >>>>decided that something other than Wintel represented a better option for >>>>servers. >>>> >>>>Then, something industry standard (and hence cheap) might emerge - which could >>>>then be used in a home PC for people who want to play GM level chess (or have >>>>high reolution virtual reality, or whatever). >>>> >>>>For many people, the best option for chess will be to rent time on a >>>>supercomputer (via the internet), rather than buy their own supercomputer just >>>>for the odd game - when a cheap computer meets all their other needs (calendar + >>>>word processor in most cases!). >>> >>>Just started a job with Sun Micro systems or the new development team >>>from intel? >> >>Not everybody does word processing. There are _plenty_ of horrendously >>complex calculations being done every day. AutoCad. Simulations. Data >>Mining. Etc. > >Data mining would normally be done on a server, rather than "personal" PC. This isn't so clear today. We have desktops with 200 gigs of disk and up. For those worried about sensitive data, using their own machine is a real option. > >Obviously if someone does animated feature films, then buying them a powerful >computer will be a cost effective investment. But for every one of those, >there'll be hundreds who use their PCs for word processing or shopping on the >internet. > >-g I think you will find that as time goes on, even the "simple" tasks are becoming compute-bound. Word processors with animated help that needs lots of horsepower. Windows itself needs horsepower to animate the windows moving and so forth. Netscape is no slouch and getting more complex all the time, same for IE. And of course, the games. :)
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.