Author: Dann Corbit
Date: 20:15:17 07/15/00
Go up one level in this thread
On July 15, 2000 at 23:13:17, ShaktiFire wrote: >On July 15, 2000 at 22:16:37, Mark Young wrote: > >>On July 15, 2000 at 19:34:18, Dann Corbit wrote: >> >>>On July 15, 2000 at 19:24:29, Ron Norris wrote: >>> >>>>What will it take to get past GM blockades? >>>>Will programmers need more knowledge, more code, cpu strength, >>>>devine intervention? What will solve this in the future? >>> >>>Three or four more plies should probably do it. >>> >>>There may be a simple algorithmic solution. >>> >>>That is not the only situation that causes problems for computers, but it's >>>probably the best one. >> >>3 or 4 more plies may be asking too much from the hardware as this amount of ply >>increase would take a huge jump in computer power. This problem could better be >>sloved with some better understanding in the program, I would think. >> >>I hear you are good at math, so I will ask...How much more computing power would >>it take for Deep Junior to see 4 more plies? > >That answer depends on the amount of pruning done, and the programs vary in >how they implement this. The simple answer looks like this: > >The effective branching factor is about 3. (Average number of "playable" >moves for a position). That means each additional ply requires a factor >of 3 in computing power. Doubling every 18 months, a factor of 3 and an >additional ply is acquired in 28 months. 4 additional plies would >require 112 months = 9 years 4 months. Moore's law has been accelerating. For the last ten years, power has doubled every year.
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.