Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!samsung!usc!jarthur!uunet!van-bc!ubc-cs!unixg.ubc.ca!cheddar.ucs.ubc.ca!buckland
From: buckland@cheddar.ucs.ubc.ca (Tony Buckland)
Newsgroups: comp.lang.fortran
Subject: Re: array addressing with pointers
Keywords: overhead
Message-ID: <1990Dec7.171857.14999@unixg.ubc.ca>
Date: 7 Dec 90 17:18:57 GMT
References: <8086@lanl.gov>
Sender: news@unixg.ubc.ca (Usenet News Account)
Reply-To: buckland@cheddar.ucs.ubc.ca (Tony Buckland)
Organization: UBC Computing Centre, Vancouver, B.C., Canada
Lines: 18

In article <8086@lanl.gov> ttw@lanl.gov (Tony Warnock) writes:
>
>      In a 1000 by 1000 by 1000 by 1000 by 50 array, would
>      someone please post the pointer structure that allows
>      access sequentially in each dimension?
>
>      How big ...

 Please tell me this is just a theoretical question.  At least
 tell me this a *very* sparse array, of which you only plan to
 store a tiny fraction of the elements, and that you're asking
 how to get from one stored element to the next stored one
 along any dimension.  Full storage of this array, assuming
 you need only fullword precision, would require 2 * 10**5
 Gigabytes, more than can conveniently be attached to most
 workstations.  About ten thousand times the size of my
 installation's mainframe disk farm.  A 300-Mby disk for every
 man, woman and child in a mid-sized U.S. city.  A lot.