Author: Vincent Vega
Date: 16:54:16 02/11/00
Go up one level in this thread
On February 11, 2000 at 19:18:02, Dann Corbit wrote: >You mean like CAP? >ftp://38.168.214.175/pub/Chess%20Analysis%20Project%20FAQ.htm No, I mean something that you can play against. Also I wouldn't start it if I expected less than about 1000 users. >If you mean a program that uses the net to play chess, then that's a lot harder >than you think. I have a very good idea how difficult it is to make chess algorithm parallel (I worked on something a bit similar). As I said, it wouldn't be real-time though. There are still tons of problems with even using alpha-beta, transposition tables, etc., etc. But that's exactly why it's so interesting. > >Does not mean that people have not tried it or are not working on it. > >But consider some of the multiple CPU efforts like Zugzwang, PConners, and >Cilkchess. Why don't they totally dominate? They have something even better >than a big batch of computers -- all the compute units are hard-wired together. The problem is the cost of making a massively parallel machine. Using the power of the Net you can bring together orders of magnitude more powerful system at a fraction of the cost. >The problem is that communication between processors needs to be very fast and >has a big overhead cost. > >These problems may be surmountable, but they are not as simple as falling off of >a log. Nobody said it would be easy. It would be challenging, interesting, and fun though.
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.