Author: chandler yergin
Date: 17:22:47 10/11/05
Go up one level in this thread
On October 11, 2005 at 19:55:12, Dann Corbit wrote:
>From here:
>http://www.albert.nu/programs/dperft/main.htm
>
>Searching the root position we have:
>Depth Perft value
>1 20
>2 400
>3 8902
>4 197281
>5 4865609
>6 119060324
>7 3195901860
>8 84998978956
>9 2439530234167
>10 69352859712417
>11 2097651003696806
>
>That means that an 11 ply search of the root position using mini/max takes:
>2,097,651,003,696,806
>That's two quadrillion nodes.
It's not an evaluation from the Starting Position,
it's an evaluation of every legal position in the game.
White has 20 Possible moves for the first move.
Black has 20 moves for the first move.
Each move and the response is considered 1 ply
The Program evaluates all the possible legal moves 1/2 ply at a time.
This is 1 iteration.
Simple as that..
;)
Chan
>
>If you can find ANY chess engine that will search the root position and get to
>11 plies I will be pretty astonished if the node count is 2097651003696806
>{meaning that every node has been examined}.
>
>On a 2GHz computer (assuming we can examine a node in one cycle which is
>probably optimistic by a factor of about 1000), that would take 1048825.5
>seconds or 12 days. Since chess engines are not just doing a perft, I would
>expect a full width 11 ply search to take a full year on the fastest AMD
>multiple CPU computer that money can buy.
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.