Robert G. Brown
rgb at phy.duke.edu
Fri Mar 1 11:02:46 PST 2002
On Fri, 1 Mar 2002, James Cownie wrote:
> > or two many controllers.
> I _think_ from what you said previously this is meant to be "too many
> controllers.", but it's possible that you're making a subtle pun and
> mean that two _is_ too many ? It might therefore be worth a correction.
Total accident -- brain fart. See previous comment about "reading
replies until 2 am":-).
Boy, I didn't expect this query to generate so much traffic! Obviously
this is a cluster design issue that a lot of people have deeply
considered and worked out a number of equally functional solutions for
(differing in cost and reliability and hence encouraging a good, healthy
ultimately cost-benefit analysis based debate:-). If it keeps up I'll
have to resummarize my summary, since Trent Piepho's posted list thread
contains results that support both sides of the hardware vs software
All in all I think the buyer beware, your mileage may vary, prototype or
test before committing to a given design rule sounds like a very sound
one, but the list archives now contain all sorts of useful starting
points and links for the cluster engineer to work with.
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