Niall O Broin wrote:
>> Anyone have any idea what that is - I mean most MIPS/FLOPS/other commonly
> accepted measure ? And how big a Beowulf would be needed to beat whatever
> that may be ? No practical application of any sort in mind, just thought
> that it might be an interesting headline :-)
>> BTW the whole idea was inspired by the University of Utrecht, in case anyone
> was curious. See http://www.lustrumcluster.uu.nl/uk/index.shtml for details.
I would suggest reading some of the articles on comp.sys.super
Theoretical Peak Performance is a very poor measure of computing
power especially when it comes to loosely coupled clusters.
For example looking at the top 500 supercomputers
No. 1: Peak = 12,288 GFlops while in tests it gets 4,938 GFlops
No. 7: Peak = 1,344 GFlops while in tests it get 1,035 GFlops
Okay some of this is the number of processors, the inter-connect
and the problem in question, so the application is VERY
important or you could go for the person with the most
boxes wins. Beowulf clusters are useful for a particular
type of problem but they have problems scaling.
Mark Fallon E-mail : Mark.Fallon at oracle.com
Senior Software Engineer Phone : +353-1-8033207
Global Product Engineering Fax : +353-1-8033221
Maintained by the ILUG website team. The aim of Linux.ie is to
support and help commercial and private users of Linux in Ireland. You can
display ILUG news in your own webpages, read backend
information to find out how. Networking services kindly provided by HEAnet, server kindly donated by
Dell. Linux is a trademark of Linus Torvalds,
used with permission. No penguins were harmed in the production or maintenance
of this highly praised website. Looking for the
Indian Linux Users' Group? Try here. If you've read all this and aren't a lawyer: you should be!