Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 11 Jul 2000 11:51:41 -0400
From:      Gary Jackson <bargle@umiacs.umd.edu>
To:        Frederik Meerwaldt <frederik@freddym.org>
Cc:        Gary Jackson <bargle@umiacs.umd.edu>, freebsd-hackers@FreeBSD.ORG
Subject:   Re: Can anyone recommend a good clustering software? 
Message-ID:  <200007111551.LAA18321@leviathan.umiacs.umd.edu>
In-Reply-To: Your message of "Tue, 11 Jul 2000 17:33:49 %2B0200." <Pine.BSF.4.21.0007111733001.1714-100000@server.wes.mee.com> 

next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 11 Jul 2000, Frederik Meerwaldt wrote:

>Hi!
>> What's PaRe?
>
>It's a clustering Software for linux.
>See:

It looks more like yet another beowulf (but built in a classroom
environment) than any coherent distribution of software.

Anyway, "clustering software" is meaningless.  What does "clustering
software" do?  Queue and run parallel jobs?  Set up a communications
medium?  Do load balancing or high availability between a set of
nodes?  All those definitions are correct.

FWIW, clustering (in the supercomputing sense) can be very simple if
you're only building a cluster for a single researcher or research
group.  You can pretty much just set up an mpi implementation, or pvm,
and an hourly passwd rdist, and call it a cluster.

MPI and PVM build on all sorts of platforms, and at least MPICH (and
probably PVM) have mechanisms for communicating between programs on
different architectures.  

Things get much more complex when you start trying to arbitrate
running time between different research units.  I won't get in to that
now, because it can be an exercise in enormous amounts of excrutiating
pain, depending on what your requirements are.

-- 
					Gary Jackson
					bargle@umiacs.umd.edu


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200007111551.LAA18321>