Author: Sune Fischer
Date: 03:40:37 08/15/03
Go up one level in this thread
On August 15, 2003 at 05:52:47, Uri Blass wrote: >On August 15, 2003 at 05:17:09, Sune Fischer wrote: > >>On August 15, 2003 at 03:02:22, Johan de Koning wrote: >> >>>On August 14, 2003 at 03:33:14, Georg v. Zimmermann wrote: >>> >>>>nt >>> >>>Thanks for appreciating my choice of words. :-) >>>There is however more to it than just fun, I wanted to avoid the B-word. >>> >>>Bug: a piece of code that does not do what the programer intended. >>>UF : a piece of code that does exactly what the programmer intended. >>> >>>... Johan >> >>I understand you point, but I think the price is too high. > >I do not know if the price is too high. The main thing that doesn't appeal to me is the thought of throwing away useful information. I guess what it comes down to is whether you want to sacrifice this to get determanistic. >> >>Clearing TTs makes the engine, IMO, significantly weaker at short time controls. >>It is simply too expensive to throw away the little information the engine has >>collected, and the fraction of a second the clearing itself takes is no small >>handicap (guess there are tricks to speed this up? :). > >Yes >You can use small tablebases for short time control. Clearing is still an overhead, even if it is less than an overhead. >Movei simply does not clear hash tables at very fast time control and starts to >clear them only if it has enough time relative to hash tables. Hmm, I think you should be consistent and either decide to clear or not. If you do clear them you get some simplicity in the hash design. -S. >Uri
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.