Author: Ernest Bonnem
Date: 11:59:36 12/10/05
Go up one level in this thread
On December 09, 2005 at 18:53:18, Ed Murak wrote: >On December 09, 2005 at 16:16:39, Ernest Bonnem wrote: > >>In http://www.talkchess.com/forums/1/message.html?467780 >>Aaron Gordon tried to look at the influence of L2 cache on the nps speed of >>Rybka. >> >>But the way the UCI launch (to depth 14) leads to numbers is strange, in all >>results reported. >> >>On my AMD mobile 2000+ (1.66 GHz, 32bit, L2=256k), the number of nodes attains >>the correct value, 3988843, and the "time" in milliseconds ranged from 53328 to >>54259 (in the 10+ tries I made). >>But my nps was always the same: 76708, >>which actually corresponds to a time of 52.000 seconds! >>Strange rounding of 53.328 to 54.259 !!! >> >>So the accuracy of the nps calculation is , to say the least, relative, >>particularly for fast machines (in the case of Aaron, the time of the 64bit >>Rybka on his fastest machine was around 15 sec) >> >>Maybe some experts have an understanding and a way to prevent such inaccuracies? > > >Yes. > >You missed my message from Dec 7, apparently - >http://www.talkchess.com/forums/1/message.html?467851 >where I already pointed this out. > >It is rounding down to whole seconds - note I do not use the term "nearest" >anywhere - and very creatively! > >So solution, also by me from the same message, is > > nps ------> dev/null > >Use instead "seconds". Hi Ed, I did see your message (which is buryed now in the depths of our CCC-Rybka Forum, so to be read I had to begin a new thread :-))), but I just wanted to dig a little further. Too bad you are so far the only one to respond! But in your message http://www.talkchess.com/forums/1/message.html?467851 you didn't say the "rounding down to whole seconds" and I can't see either your "nps ------> dev/null". Did I miss something? And more importantly, what do you mean by: solution... nps ------> dev/null? The "granularity" you mention is clear, since the nps number is obtained by dividing 3988843 by some "round" (integer) seconds (as low as 15 for Aaron, 52 for my slow machine). But I am also wondering why, by repeating my tries (more than 10), I obtain time values ranging from 53.328 to 54.259. This reminds me of the granularity of "timer" values in DOS/Basic (also Windows?), of 1/18th of a second (55 msec).
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.