Author: Robert Hyatt
Date: 11:12:46 10/18/04
Go up one level in this thread
On October 18, 2004 at 11:40:55, GeoffW wrote: >Hi Bob > >Thanks for the help, just trying to digest some of your other points, but this >one just made me smile in relation to you first comment > >>>I am using SEE to order the capture moves in my Q search. >>>Move ordering in the normal search is still MVVLVA as my SEE is quite slow >> >>This is backward thinking. Look at how many qsearch nodes there are vs the >>non-qsearch nodes. Yet you are using your "slow SEE" where you use it far more >>often than if you used in in non-qsearch only. :) > >I was just testing a position with and without SEE changes > >I do keep a count of nodes and qnodes, but must admit I havent really studied >the relative percentages very much so dont have a very good feel for it. >I then looked at the stats output line > >11+ 15 2146 26800500 f4c7 >11+ 315 2159 26959969 f4c7 >11 9977 5979 92721403 f4c7 c8c7 c3b5 c7b8 d1d8 e8d8 e3f4 g4e5 f4e5 b8a8 >b5c7 a8b8 c7a6 b8a8 a6c7 a8b8 c7d5 d8d6 e5d6 b8a8 d5b6 a7b6 c1a1 > >Computer's move: f4c7 > >Nodes=92,721,403 Quiescent Nodes=48,694,293 Qnodes Percent=6.0% > >Note my very low Qnodes percentage, and the slight rounding error. Cough !! > >Fixed that now, silly 32 bit overflow problem, but I have even less feel now for >typical ratios of qnodes/nodes. > >I guess the point you are making is that qnodes are large in relation to nodes. > > Geoff Correct. IE your first call to quiesce() is mandatory, and there will be as many "first calls" as there are total non-q-search nodes roughly. IE you spend way more time in the q-search than in the normal search. That has to be where you worry about efficiency...
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.