[Beowulf] commercial clusters

Stu Midgley sdm900 at gmail.com
Fri Sep 29 08:58:58 PDT 2006


??  NFS?  And you claim performance is a problem?  I'm not surprised.

Of the machines I use and manage, I happily see 450MB/s (168 cpu
machine) and 2.5GB/s (1936 cpu machine)...




>      I too work for one of those Houston companies and in our cluster performance absolutely does count.  The way it is structured here, we compete with outside businesses for BU jobs, so we have to perform.  I like to use "bigger-faster-cheaper" as our slogan for production work.   And among our bottlenecks is network and nfs performance.   We push nfs so hard we have to tweak file systems to make it handle the work. Because right behind that job are tons of more jobs with people waiting get their jobs done so their projects can move forward.
>
>      Our next generation software will absolutely beat the snot out of our existing storage (and I pride myself on performance). So with another 100TB coming in the next few months we are always configuring and testing.  I have taken a left turn for a bit to explore how using grid technology to increase our SLAs, but I will get back to exploring options to make our system more efficient.



-- 
Dr Stuart Midgley
sdm900 at gmail.com



More information about the Beowulf mailing list