Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!samsung!sol.ctr.columbia.edu!bronze!mosque.huji.ac.il!ury
From: ury@mosque.huji.ac.il (ury segal)
Newsgroups: comp.sys.super
Subject: Paralel supres    vs.    SISD supers
Message-ID: <1991Jun3.141024.14991@bronze.ucs.indiana.edu>
Date: 3 Jun 91 14:10:24 GMT
Sender: news@bronze.ucs.indiana.edu (USENET News System)
Organization: Hebrew University of Jerusalem, Israel
Lines: 13


How far can we go with SISD (Signle instraction, signle data) machines ?
every given computer is not fast enough, by defenition. So paralel and/or
distributed computers is the answer, but each node will cost too much if it
will be a super (Think about MOSIX run on 64 vax 9000 ....) . So they make
every node from a fast CPU (i860) but not the fastest.
Think about machine with 64 nodes, each as fast as cray 2.
Think about the almost-unlimited speed  (Will be enough for us ??? )
( Think about the *COST*    ;-)  )



--ury