/. MS vs. university Beowulfs

Robert G. Brown rgb at phy.duke.edu
Fri Feb 22 08:49:00 PST 2002

On Fri, 22 Feb 2002, Velocet wrote:

> Have fun wading through the blind linux advocacy and silly jokes to try
> and find any content.
> If someone can summarize any real information they grok from this massive
> thread (ie how good is computational clustering on M$), please post it here.

Or not posting it here is OK too, recalling that the word "free" appears
explicitly in the FAQ "What's a Beowulf" (as maintained by Kragen
Sitaker) and "open source" (e.g. Gnu and friends, especially Linux,
OpenBSD etc) is very clearly evident in the philosophy underlying the
beowulf design as put forth on at beowulf.org.  

M$ is not free, nor is it open, nor is it secure (I'm getting WinXX
viruses on a daily basis now -- the frequency is actually competing with
that of SPAM!) nor does it work particularly well, last time I tried it.
The relatively inexpensive WinXX's are operating systems where you have
to reboot to restart or reconfigure the network.  The somewhat more
posix-compliant and robust WinNT derivatives are hideously expensive --
they can damn near double your per-node costs all by themselves, by the
time you get them fully equipped with compilers and so forth.  Nodes
need to be stable -- months of flawless uptime under load is not too
much -- and M$ operating environments are not exactly world-reknowned
for their operational stability.  And of course I could go on -- this is
a perrennial discussion on the list with plenty of content in the

Still, I wouldn't mind a real >>summary<< (as in short and to the point)
of current clustering state-of-the-art on e.g. W2K or WXP clusters, but
please, for the love of God Montressor, not a recurrence of the usual
mix of MS marketing bullshit and do NOT call a M$ cluster "a beowulf" as
it isn't...;-)


Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu

More information about the Beowulf mailing list