Author: Peter Fendrich
Date: 04:02:43 12/30/01
Go up one level in this thread
On December 29, 2001 at 20:24:10, Uri Blass wrote: >On December 29, 2001 at 20:09:15, Peter Fendrich wrote: > >>On December 29, 2001 at 19:13:33, Christophe Theron wrote: >> >>>On December 29, 2001 at 18:48:32, Severi Salminen wrote: >>> >>>>>>>If your full QS eats 80% of the time I would say that limiting QS is indeed a >>>>>>>must. However my QS (full) takes about 10-15% (25-30% with checks) and so the >>>>>>>picture is quite different then. >>>>>> >>>>>>Ok, but what about nodes? Does your qsearch take only 10-15% of all the nodes? >>>>>>That's quite low... >>>>>> >>>>>>Severi >>>>> >>>>>Yes that is very impressive, I must have lots of room for improvement ;) >>>> >>>>Yeah, but no way it can be 10% of all the nodes: that is just too low. Or then >>>>it prunes like hell... >>>> >>>>Severi >>> >>> >>> >>>Believe it or not, above 30% you have a big problem somewhere. >>> >>>That's the reason why using a SEE instead of the QSearch is not an improvement. >>> >>>Remember: we do not count the horizon nodes as being in the QSearch, because >>>with or without QSearch you have to visit them anyway. >>> >>> >>> >>> Christophe >> >>This is quite funny... >>It's not the first time where there is a long thread about qsearch performance >>and it turns out that people count both nodes and Qnodes different. >>Especially the transition from ordinary search to Qsearch seems to be handled >>differently. >>What exactly do you mean with horizon nodes? >> >>//Peter > >It is clear. > >horizon nodes are the moves that you make before calling qsearch. > >you could not avoid making them even without qsearch so you should not count >them. > >after 1.e4 e5 2.d4 the qsearch incluse exactly 2 nodes >2...exd4 3.Qxd4 and not 3 nodes(2.d4 exd4 3.Qxd4) > >Uri If it's that clear, why did Christofe have to point out that he doesn't count 2.d4? Of cource he shouldn't count that. There are so many ways of implementations and there are grey zones about how to count and I'm sure there are also misunderstandings. One example: Take a look at Bruce M's implementation of QSearch that detects checks. http://www.seanet.com/~brucemo/topics/quiescent.htm When in check in Qsearch, he calls AlphaBeta(...) How will it be counted? There are number of others as well. I'm not trying to say that this is an errror and of cource it is possible to catch that in the code, but we have a communication error when talking about x% Qnodes. //Peter
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.