Beowulf usage statistics?

[Roger L. Smith]

On Mon, 24 Jun 2002, Velocet wrote:

> Perhaps people have designed their clusters for actual use instead of
> benchmarking to get into the rankings of the top N in the world, realising
> their clusters wont perform well in that type of ranking.

Well, even some clusters that aren't so well suited to Linpack
benchmarking show reasonably well in the top 500.

> Its just a testostermoronic size competition IMHO.

Yes, but it's also sometimes a useful marketing tool.  Our university
administration is quite fond of touting our position on the Top 500 list
as a tool to help recruit new students, and to prove their dedication to
our computation research efforts to prospective new rsearch partners.

We all know that the numbers are nearly meaningless, and I would certainly
have built the system differently if that were my only goal, but it is
certainly a useful piece of propaganda to have under my belt in certain
situations.


 _\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_\|/_
| Roger L. Smith                        Phone: 662-325-3625               |
| Systems Administrator                 FAX:   662-325-7692               |
| roger at ERC.MsState.Edu                 http://WWW.ERC.MsState.Edu/~roger |
|                       Mississippi State University                      |
|_______________________Engineering Research Center_______________________|