Author: Ricardo Gibert
Date: 19:35:34 07/03/03
Go up one level in this thread
On July 03, 2003 at 22:24:01, Dann Corbit wrote: >On July 03, 2003 at 22:12:03, Mick Turner wrote: > >>If I purchased a box with 10.20 ghx X 8cpu; >> >>what type of calculation speed would I be getting ? > >It does not exist, so any answer would be hypothetical. However, let's suppose >that a 1GHz 64 bit chip can do a million chess calculations per second, and just >for fun we will scale a 10GHz machine to do 10 million. Then if we have 0% SMP >loss, you would have 80 M NPS. Deep Blue sustained 200 M NPS. Of course, all >of this is meaningless. There are slow searchers that might play better and get 1 billion NPS was its theoretical peak 200 million NPS was a kind of practical peak 125 million NPS was its normal rate Hsu himself caused the confusion of the 200 million NPS figure with the 125 million NPS figure, but he was pinned down on what was what a while ago in a Q&A session on ICC. >1/10 of the NPS. Or we might rip out the eval -- count only wood -- and do one >trillion NPS, but get killed with zero points in a thousand game match. > >>How would it compare to DB hardware ? > >It would be either weaker, or stronger. Unless it was the same. Trichotomy and >all of that.
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.