Author: Robert Hyatt
Date: 10:46:20 07/13/99
Go up one level in this thread
On July 13, 1999 at 13:26:45, Eugene Nalimov wrote: >On July 13, 1999 at 04:00:04, Shep wrote: > >>On July 12, 1999 at 20:30:06, Gustavo Pereira wrote: >> >> >>>Remember that node count is not a very valid way to account for a program's >>>strength, take LGG2.05 as an example, I have seen it speeding far away from >>>Fritz, yet it is not a very good chess playing program, it is an excellent >>>problem (tactics!!) solver. >> >>LGG2 is probably the program with the most differing nps counts of all. >>I have seen it hit anywhere between 200 K and 1,300 K on the PIII-550. >>(No other program I know differs by a factor of more than 3.) >>Whenever there are tactical shots in the position, LGG's node count explodes >>whereas it remains low in positional situations. >>Is there any explanation on why this happens? >> >>--- >>Shep > >Probably in tactical positions it uses fast evaluation function (e.g. material >only, or some form of piece square tables), while in "normal" positions it uses >slow eval. A lot of chess programs do the same, but probably in that case the >difference in speed is greater than "average". > >Eugene Good point...for those doing 'lazy evaluation'... forgot about that aspect of speeding up.
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.