Uses for a beowulf cluster?

[Daniel J. Frasnelli]

Greetings,

On Sun, 13 Sep 1998, Gerry Creager wrote:
> What you're really describing is a small (or maybe, large) network of
> workstations... and I don't think you're referring to a NOW.  If all are
> running LINUX, have NFS and NIS and such, they will accomplish what
	Right, but I do not believe that you can meet these requirements 
"out of the box" with any given Linux distribution.  A network of
workstations (or servers) without underlying communication glue is just a
group of computers connected to a network.  To exploit the potential for
parallel computation, you must provide an adequate vector for
communication between the nodes -- this is why I recommended a "load
sharing" scheme rather than trying to develop a distributed userland
environment.    
> you've outlines, with the added benefit that they *CAN* be used in a
> scalable parallel configuration if needed.  As already indicated, the
> "cluster" we usually consider a "Beowulf cluseter" appears as a single
> parallel entity, although it may in reality be a distributed lab of
	Sure, any group of networked computers with the proper software
can work as a parallel machine.  I'm using a lab of 30 Sun SS5/170's, and
am quite satisfied with the results.  
> computers, all running a coherent operating system and residing (for
> discussions' sake) on a single isolated network, or might be a bunch of
> computers doing nothing but parallel computation stuch in a rack or
> wiring closet somewhere.
> Preplanning for multiple use is a good idea, and I applaud you on it.
	One of my recommended solutions for the department's dilemma 
is to purchase four 600Mhz 21164a's (4M L3 cache) with 1024M ram and a
local 4.5G UW-SCSI drive set up using the load sharing mechanism outlined
in a previous message (Each node is approx. $10k).  This, of course, is a
subtle way of obtaining a platform for parallel computation: Four very
speedy processors with 4096M of shared memory.  (BTW, this is quite
tentative and in the future, so don't get too excited 8^)

Best regards,
Daniel  

---
Daniel J. Frasnelli		Remote sensing scientist
dfrasnel @ wvu.edu		Imaging spectroscopy researcher
Explore terrestrial physics! http://ltpwww.gsfc.nasa.gov/