Author: Charles Roberson
Date: 20:32:11 11/08/02
Go up one level in this thread
When multithreading tree/graph search algorithms, each thread do not need to execute the same instructions in lock synch. Thus, reducing the issue. Also, one can make most of the data local in scope to the search function. Thus, in a recursive search, each depth has its own stack frame. This works well with multi-threading as each thread has its own stack. This can make your stall a completely non-issue. This method is quite common and does apply to chess. Charles Roberson
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.