Author: Robert Hyatt
Date: 08:28:21 04/12/04
Go up one level in this thread
On April 11, 2004 at 07:46:59, Tord Romstad wrote: >On April 10, 2004 at 16:58:07, Dan Andersson wrote: > >> A wild guess is, 'Dynamic Tree Splitting (DTS)'. Not to be confused with TDS :) >> There is an article on Bobs site: http://www.cis.uab.edu/hyatt/search.html > >Looks like a good guess. Thanks, Dan! > >If it is not possible to write a parallel search algorithm without a >non-recursive alpha >beta function, I am definitely not going to do it. :-) > >Tord Crafty is recursive negamax and the parallel search works just fine. There is a problem caused by the recursive calls, but as you can tell from Crafty, it is solvable... There is nothing wrong with non-recursive negamax either. That is what I did in Cray Blitz, but it was done not as an intentional design issue to make DTS easier, it was done because early FORTRAN didn't support recursion... Non-recursive has an advantage in that you can split anywhere in the current path with no difficulty, while to do this in a recursive search you have to do some work...
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.