Author: Ricardo Gibert
Date: 21:22:01 05/28/02
Go up one level in this thread
On May 28, 2002 at 23:47:11, Ricardo Gibert wrote: >On May 28, 2002 at 22:28:32, Russell Reagan wrote: > >>On May 28, 2002 at 17:18:37, J. Wesley Cleveland wrote: >> >>>Currently, processor speed and hard disk size are also increasing exponentially. >>>As long as this continues, we will be able to generate larger and larger >>>tablebases. >> >>If this is the case, then it's obviously growing at a much lower exponential >>rate, since the 5-man tablebases can be generated in a matter of hours and will >>fit on almost any new computer's hard drive, and the 6-man tablebase will take >>several years to complete and will probably not fit on any new computer's hard >>drive even in a few years when they are completed. >> >>So the problem still exists. The rate of growth of the size and space used by >>the tablebases is increasing at a much higher rate than the speed and storage >>capacity of computers, so the problem still exists regardless of whether both >>factors are growing at an exponential rate. >> >>Russell > >Even though the exponential rates are different the ratios of the rates between >hard disk space, cpu speed, tablebase size are constants as long as the model of >exponential growth holds in each case e.g. if item X grows at rate a^t and item >Y grows at b^t where t is time, then a^t/b^t = a/b = some constant. > >You're comparing the time it takes to generate an EGTB on the same hardware when >the capabilities of hardware actually vary with time in a exponential way for >the forseeable future. > >[J]WC actually made an astute point. I would guesstimate it takes about 9 years to advance from the N-man EGTBs to the (N+1)-man EGTBs for as long as the exponential model holds in all relevant areas.
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.