Author: Frank Schneider
Date: 06:21:38 06/01/98
Has someone tried to parallelize a program like described below? It seems quite simple and could be easy to implement, but I'm not sure if it would give good speedups. Since it might be an easy way to add parallelization to a sequential program I'd like to know if it works. a. One process performs a (sequential) alpha-beta-search b. other processes help and provide results in a shared transposition-table (tablebase-access must be synchronized) There are several possibilities for b.: - other processes scan the transposition-table and try to improve stored results(replace bounds by values or by bounds/values returned by a search to greater depth) - other processes are started by a. For example whenever a. searches a positon to depth d (evaluation is d=0), another process starts to search the position to depth d+1 and stores the results in the transposition-table. - ... Frank
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.