Author: Robert Hyatt
Date: 20:14:38 08/19/00
Go up one level in this thread
On August 19, 2000 at 15:49:06, tony worsman wrote: >Hi All, > >When testing Crafty on a position using long time controls, I've noticed that >the node counter reaches 4-5 billion, but never goes beyond this number, but >decreases giving a false nps. As the search does not appear to be affected >by this behaviour, It is not really a problem. > >Is this a limitation of 32 bit computers, Crafty, or the C language, not sure, >I'd be interested to know the answer. It is a 32 bit problem. I use an unsigned long, which can count to just over 4B nodes, then it wraps back around to zero. > >Here is Craftys output on the same position, the first is 333 min, the second >is 500 min > > > > time=333:20 cpu=99% mat=-1 n=4160976338 fh=10% nps=208044 > ext-> checks=209585241 recaps=13647126 pawns=4243489 1rep=19302187 >thrt:1356639 > predicted=0 nodes=4160976338 evals=827478374 > endgame tablebase-> probes done=0 successful=0 > > > > time=500:00 cpu=99% mat=-1 n=1837858478 fh=8% nps=61261 > ext-> checks=345064260 recaps=20089204 pawns=10840711 >1rep=30816560 thrt:2019519 > predicted=0 nodes=1837858478 evals=1180975979 > endgame tablebase-> probes done=0 successful=0 > > >Thanks........Tony.
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.