[comp.realtime] Reference wanted: Multiprocessor Scheduling

bigm@batserver.cs.uq.oz.au (Michael Pilling (local Big M Sella Fella)) (11/15/90)

Hi, I'm after a reference (preferably the earliest) that explains why
it is better to schedule processes on a multiprocessor individually
rather than treating it as a processor pool and running the top
n processes on the multiprocessor (i.e. global scheduling).

I suspect the reason is that static process placement that is
near to harmonic as possible is optimal, but am after a proof as to why.

Michael
--
 _--_|\  Michael John Pilling,		 bigm@batserver.cs.uq.oz.au
/      \ Computer Science Department,    She was pregnant and, unfortunately,
\_.--._/ University of Queensland,	 the serious bruise to her back
      v  Queensland, 4072 Australia.     prompted her to abort.[Daily Sun]