Author: Robert Hyatt
Date: 18:13:01 02/17/05
Go up one level in this thread
On February 17, 2005 at 19:27:51, Dieter Buerssner wrote: >On February 17, 2005 at 16:27:06, Robert Hyatt wrote: > >>For example, none of the EGTB code will be profiled >>if you don't use a couple of endgames. > >It won't matter. 90% in EGTB accessing will be used for fread(). There is not >much potential (one could say practically no potential) for PGO for EGTB access. >At least not in positions I tried and rather carefully profiled (this included >positions, where over 90% of the ovarall time was used for TB probing). Perhaps >with really fast disks, the situation is slightly different, but still, the >potential for PGO will be very small here. > >Regards, >Dieter I've got "really fast disks" here. U320 15K scsi drives. But I don't have all endgame tables on my machine, which means I call EGTBProbe() many times and it goes through its internal checks and returns failure. Those can be optimized... Plus all the special-case endgame evaluation that would never be touched from an opening position, but which would be important in an endgame position later in the game...
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.