Author: Vincent Diepeveen
Date: 08:19:05 05/25/01
Go up one level in this thread
On May 24, 2001 at 15:47:04, K. Burcham wrote: Xeon is not AMD but intel. dual 1.7Ghz Xeon/P4 is going to be a big dissappointment for any person who tries a chessprogram at it! The faster the processor the bigger the problems of the small L1 cache will be not to mention the 20 clocks penalty for branch mispredictions and the lack to compensate it. bandwidth became better but that's no issue for computerchess, for quake it is however. > >with fritz6 128 megs hash > >16 mhz fidelity = who cares > >33 mhz intel = who cares > >550 mhz pentium lll = 495 kns > >1000 mhz pentium lll = 900 kns > >1000 mhz amd t-bird = 900 kns > >1500 mhz amd t-bird = 1350 kns > >dual 750 pentium lll = 1180 kns > > > >1400 mhz pentium llll = 1050 kns > >1700 mhz pentium llll = 1275 kns > > >dual 1700 pentium llll = 2210 kns > >dual 1700 amd = 2750 kns > >i would not be surprised if the dual amd hits 2900 kns, i was conservetive >with the dual amd numbers. > > > >the first eight computers i own or have owned. I ran the fritzmark test >to get these numbers. all used 128 megs hash. > >the pentium llll's, I base the kns on the cpu benchtests i have seen at >the internet sites. yes i am aware that the software used was not a chess >program. but i think we will find that these numbers are close.
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.