Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uunet!ncrlnk!ncrcae!hubcap!wen-king
From: wen-king@csvax.caltech.edu (Wen-King Su)
Newsgroups: comp.parallel
Subject: Re: scalability of n-cubes, meshes (was: IPSC Communications)
Message-ID: <7487@hubcap.clemson.edu>
Date: 19 Dec 89 14:22:42 GMT
Sender: fpst@hubcap.clemson.edu
Lines: 23
Approved: parallel@hubcap.clemson.edu

<From: cleary@cpsc.ucalgary.ca (John Cleary)
>In article <7427@hubcap.clemson.edu>, landman@hanami.Eng.Sun.COM (Howard A. Landman x61391) writes:
<> >In article <7274@hubcap.clemson.edu> gene@cs.bu.edu (Gene Itkis) writes:
>> >%One problem with 3D mesh is that as the processors work the heat is
<> >%generated(*).....
>> 
<It worth noting however that the temperature of a 3D cube rises very slowly
>with the number of nodes.  The heat lost by radiation (probably the only
<feasible mechanism in the limit) is porportional to A*T^4 where A is the
>surface area and T is the temperature.  If the radius of the 3D body is R then
<A scales as R^2 and the number of processors (N) and hence the total heat (per
>unit time) as R^3.  The result is that T scales as N^(1/12) pretty slow.

'T' is the temperature at the surface of the 3D cube.  In order to allow
heat to dessipate, there must be a thermal gradiant from the core to the
surface.  In order to keep the core temperature at an acceptiable level,
the surface temperature must drop (not rise, nor to remain constant) as
R is increased.  At a certain R, the required surface temperature will
hit 0 degree Kelvin, and we can go no further.

/*------------------------------------------------------------------------*\
| Wen-King Su  wen-king@vlsi.caltech.edu  Caltech Corp of Cosmic Engineers |
\*------------------------------------------------------------------------*/