Author: Tom Kerrigan
Date: 16:04:59 05/29/00
Go up one level in this thread
The algorithm is not extremely easy to understand from those three rules. Here's what happens. Imagine 2 threads doing a 1 ply search: * Both threads start by searching the highest-ordered move. This sets alpha in both threads. * Now the goal is to find moves that are better than alpha. Thread 1 starts searching move 2. Thread 2 tries to search move 2, but it's already being searched, so thread 2 skips it and searches move 3. This game of leapfrog continues until each thread has searched half the moves. * The threads go back to search the moves they skipped. But the skipped moves have already been searched by the other thread, so "searching" them is basically free (because of the hash table). You can imagine how this algorithm breaks down if move ordering isn't very good. Fortunately move ordering in chess is excellent. Hope this helps. -Tom
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.