Author: Robert Hyatt
Date: 07:56:30 01/18/98
Go up one level in this thread
On January 18, 1998 at 09:47:40, Ernst A. Heinz wrote: >On January 18, 1998 at 07:36:26, Peter W. Gillgasch wrote: > >>00:01:00 6122978 nodes NPS: 101710.598007 > >Hi Peter, > >in terms of NPS "Chess Demon" seems to be hellishly fast indeed -- >congratulations! > >But what happened to your tree size? The current "DarkThought" needs >only 3.8 million nodes to complete iteration #11 of WAC #3. > >=Ernst= Could be explained by lots of things... 1. different search extensions. 2. different evaluations. IE, on occasion I will slow down Crafty solving one of the WAC positions just by making an eval change that affects the size of the tree because it suddenly makes a branch look better than it did before, but, in the long-run, not better enough to change anything. I need 5.3 M nodes to search 11 plies, although I find it at ply=6 in less than .1 seconds.. We could compare better if we took the node count for the entire suite, to a fixed depth, and compare the sums. That would tend to knock out minor eval differences. But search extensions can be important. I'm trying to reduce mine a bit now. When someone posted that mate in 30 position and I found Crafty could solve it, I immediately started working on *not* solving it. There is no sense being out at ply=60 with a full-width search, IMHO.
This page took 0.01 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.