Author: Russell Reagan
Date: 09:10:08 08/18/03
Go up one level in this thread
On August 18, 2003 at 11:39:08, Robert Hyatt wrote: >Anything is possible, of course. But you won't just lose "some >speed". You will lose a _bunch_. This is a well-understood problem. Getting a bit more off topic here, but you could achieve an Nx speedup on N processors using min-max, right? With pure min-max this wouldn't work well in practice, but maybe some kind of selective min-max could be competitve with alpha-beta based algorithsm as SMP/HT machines become more and more affordable. The selective min-max could still make use of alpha-beta/pvs/mtd/etc. My thinking is that this wouldn't pay off great dividends immediately, but eventually we will probably run into some kind of barrier as to new improvements to a single cpu's speed. One obvious continuation is for the cpu creators to start piling single cpus together as they've already started doing with HT. If in 30/40/50+ years we have a cpu that has a million cpus within it, we might see the payoffs there.
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.