Author: Chris Carson
Date: 07:21:49 08/06/99
Go up one level in this thread
On August 06, 1999 at 09:24:28, Randy Schmidt wrote: >Thank you for the explanation! > >Now the hard question... > >If a program has a 2550 SSDF rating, what would the estimate >be for 'strength' increase by using a Xeon vs. a regular P3. >We can assume the same 400 MHz, etc. This is difficult to predict. MHZ is not a reliable measure of increase in processing power. SPECINT is better but not exact. :) Most programs are optimized for a certain mp family (PII, PIII, Xeon, ...) and may not perform as well on other mp families. Another factor is that some programs gain more from mp/memory increase than others. So the only reliable way to tell is to test your program on the target platform against other computers/people. Just my opinion and observation. With Crafty you can run the bench and get a good estimate of NPS and compare the mp you are on. If you get a speed increase of 2x on your faster machine then you may see a SSDF increase in ratings between 30 and 60 points, however the error of measure is in the range of 25 points, so the estimate is not reliable, only extensive testing can establish a reliable rating. Just my opinion and I am not an expert. :) Best Regards, Chris Carson
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.