Author: Ingo Lindam
Date: 06:02:25 10/28/02
Go up one level in this thread
On October 27, 2002 at 20:21:34, Bob Durrett wrote: >Similarly, in the middle of [in the gory guts of] a search algorithm, maybe >some of the positions occuring during the search could be evaluated the same >way. In >a purely serial machine [no parallel processing] I fear that the time required >for this computation might not leave enough time for the rest of the search. In >other words, "not competitive." If you are going serial, you have a big >challenge to make your engine competitive with current engines which do not [?] >use your idea. Parallel processing is another matter, given the requisite >technology. Perhaps ... or perhaps not... ofcourse I would appreciate to divide the work of evaluation on a lot of parallel processors and the approach gives ofcourse some natural oportunities to divide it. But unfortunately I don't have access on a massive parallel system yet... so I have to watch how far I may come without it. An advantage of the approach might be to cut off a lot of the big tree and then have to compute much less nodes by using the knowledge. >Am I still at least "out in left field" on this one? [Still in the ballpark?] Yes, you should still be in the ball park... because I am here in the infield catching your balls you throw towards me. >You envision producing a "black box" with inputs and outputs. The inputs would >consist of one or two million master level chess games. The outputs would be a >large set of patterns with associated properties &/or other useful data. > >Right? [If yes, then how?] Imagine... (hear the music) Imagine, you spend a little more storage to represent the positions of the 2 million games that might be about 100 million positions in a data structure that allows you to have an efficient access on the information in which positions a pattern occurs. Now assume a and b to be pattern for wich you already know in which positions they occure. Then it is a very easy and efficient to obtain question in which positions the pattern c1 = a AND b and in which the pattern c2 = a OR b occur. This can be obtained very very fast. And when you now further can use the fact that in chess you may a very unequal number of 0s and 1s in this representation you might save again a lot of storage and time. Ingo
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.