Author: Tord Romstad
Date: 00:44:15 12/06/03
Go up one level in this thread
On December 05, 2003 at 18:49:37, Matthew McKnight wrote: >On December 05, 2003 at 18:31:34, Matthias Gemuh wrote: > >> >> >> >>The ratio nullmove_cutoffs / main_search_moves_searched is only 15%...20% >>in my program. Is this extremely low ? I am not quite sure I understand what you are measuring here. Is main_search_moves_searched the total number of non-qsearch nodes, or something different? >180 kN/s is not bad if you are doing a lot of work in eval, or other things that >slow you down, as long as you are compensated. With BigLion's performance, I'd >say you are succeeding here. Agreed on both counts. 180 kN/s is fine. Gothmog is not much faster (at least not in the middle game, but it speeds up more in the endgame than your engines), and I expect the new engine I am writing to be even slower when it is finished. The main weakness of BigLion and Taktix seems to be their search. With your big and high-quality eval, you should be able to search about 2 plies deeper in the middle game by using the eval to do forward pruning and/or depth reductions. One thing I have noticed is that your engines' behaviour in the first few iterations is rather unusual. In almost all other engines, the first 4 or 5 plies flies by so quickly that it is not possible to follow the lines on the computer screen. In BigLion and Taktix, each of the first 5 plies take about a quarter of a second to complete. This is strange in two different ways: It is strange that a 1-ply search should take as much as a quarter of a second, and it is strange that the amount of time needed is apparently almost linear (rather than exponential) as a function of the search depth. For bigger search depths, things begin to look more normal. What's happening in the first few iterations of your search? Tord
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.