Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site ucbvax.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!ucbvax!isi-vaxa.arpa!lerner
From: lerner@ISI-VAXA.ARPA (Mitchell Lerner)
Newsgroups: fa.info-vax
Subject: Re: Cluster overhead
Message-ID: <8509111730.AA16997@isi-vaxa.ARPA>
Date: Wed, 11-Sep-85 13:30:00 EDT
Article-I.D.: isi-vaxa.8509111730.AA16997
Posted: Wed Sep 11 13:30:00 1985
Date-Received: Thu, 12-Sep-85 11:55:13 EDT
References: <8509110225.AA08696@isi-vaxa.ARPA>
Sender: daemon@ucbvax.ARPA
Reply-To: info-vax@ucb-vax.arpa
Organization: The ARPA Internet
Lines: 16


Dear Mike:

	Its not as simple as DEC makes it sound in the cluster literature.
Many beings have told me that they have built a clustered system and 
jobs ran actually slower due to a non optimized cluster.  From what I 
understand about clustering (which isnt much), one needs to hone everything
on the cluster to a fine edge inorder to realize any performance gains.
"What the "tweakable" variable are exactly and just how to tweak them
is a job only a "DEC cluster consultant" can tell us", I am told.

Good luck!!!

							Sincerely,

							Mitchell