Author: KarinsDad
Date: 23:11:28 02/02/99
Go up one level in this thread
On February 03, 1999 at 00:53:52, Dann Corbit wrote: >On February 03, 1999 at 00:37:47, Matt Frank wrote: >[snip] >>Deeper Blue (circa 1997 match) hit about 200,000,000 nodes per sec. If this new >>monster you speak of could hit 2,000,000,000 per sec. And they could port the >>same software to this new creature, with no significant loss of efficiency, the >>new (Deepest Blue :-}) would be rated at least 200 elo points higher than it's >>1997 predecessor. Formidable indeed!! >Number mixing here is apples and oranges. Two billion operations per second is >not the same thing as two billion chess nodes per second. Look at Crafty -- the >NPS figure is far different from the CPU MHz rate, and Crafty is a fast >searcher. If they have updated the Chess CPU chips to get two billion chess >positions per second, then that is another story. The updated CPU's and I/O >bandwidth would improve the machine, certainly. But not in the manner that some >may be thinking. The original posting here says 2x faster (which should extrapolate to 400M nps). However, I read something a few weeks back on the IBM web page that said the new chips from last summer were 5x faster (i.e. 1B nps). However, 5x seems like a fairly major jump in one year. I would tend to think that the 2x number is more accurate. Someone should look it up on the IBM web page. It's to late at night for me to bother. KarinsDad
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.