Author: Graham Laight
Date: 01:30:23 08/03/01
Go up one level in this thread
On August 02, 2001 at 15:07:44, Bruce Moreland wrote: >I think the best thing is if they say SMP is the way to go, we're going to >provide them next year, so shut up and get with it now. > >Then people would have a year to sink or swim. In the long term, it is inevitable that, to keep the standard of chess rising, multiprocessor machines will have to be admitted (IMO). >Of course they'd never find a sponsor who'd contribute 24 duals, but the writing >was on the wall the moment Crafty showed up on an SMP machine on ICC. > >If I actually get on the plane rather than tearing up my ticket, the version I >use will be my one and only version. If you put it on an 8-way, it will use all >8 processors. It will actually run on a lot more, if a lot more are available, >but I don't think there are any systems that support more. If you put it on a >single, it will use one. That's the way to do it. > >bruce The winners will be teams whose programs are portable between platforms - and who can persuade Unix companies to lend them their latest models of 64 bit machines with huge numbers of processors. Last year, I persuaded Compaq to lend me a (single processor) 64 bit Alpha to take to Alexandra Palace to run Crafty. However, it turned out to be an old model, and I didn't get it soon enough - so I had to spend all my available time just getting it to work properly. We only got half the NPS we would have been able to get on a new Alpha. The losers of course will be Fritz. Writing their program in assembly may have kept them at the top of the hill for a long time - but it's looking like a serious handicap now. -g
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.