Author: Will Singleton
Date: 17:54:04 07/26/99
Go up one level in this thread
On July 26, 1999 at 17:54:03, Dan Newman wrote: >Result for Dann's executable on P6/200, depth=7: >time: 60.750000 sec., nodes=1557668, NPS = 25640.625514 > >I typically get a rise in node rate with the length of >the test (in this case 23k -> 25k). I've always attributed >this to a "settling in" effect due to the branch prediction >hardware and perhaps some sort of cache effect. Any thoughts? > >-Dan. I might mention that the nps results are only valid from the start of program execution. That is, if you do "on new on", you'll get different results (typically slower, but fewer nodes) the second run. Have no idea why this is, unless he doesn't clear his history tables or other data on new game. Will
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.