On Thu, 11 Nov 2004 15:24:37 +1100
"Nigel Weeks" <nweeks@examiner.com.au> wrote:
> As soon as the load average reaches 1000, the load then drops away to 20 or
> so, then climbs back up to 1000
> Is this like a 'rev limiter' for the scheduler?
At a guess, sounds more like some sort of Nyquist-like effect.
http://foldoc.doc.ic.ac.uk/foldoc/foldoc.cgi?Nyquist+Theorem
-Chris