[Bioclusters] cluster or SMP

[Steve Pittard]

Joe Landman said:
>
> We have built desktop units that have 4-8 cores, 32 GB ram, and >1TB
> disk with 500+ MB/s bandwidth for customers running research apps (CFD,
> bio, chem, ...) .  With the soon to be available Barcelona, we will be
> able to pack 16 cores, 64 GB ram and a fast IO subsystem in a single
> system image deskside (SMP).
>
> We are seeing *lots* of interest in these systems, including people
> wanting to build clusters out of them.

This is something that I've been seeing at a local level.
Many would-be users of a larger cluster resource typically want some
platform for job preparation as well as for post cluster run analysis.
While some users can accomplish this successfully on a beefy desktop
others require a more substantital setup with 16-32GB of RAM perhaps
to support a RDBMS and a computational environment. Users who manage dense
data sets (such as those associated with genomic projects) actually need
to run sets of follow on jobs to distill their candidate results into
something less granular and more manageable.

So the SMP boxes are helping in this area. Of course we can still use
SGE to manage work on these setups but its more common to find departments
(at least where I am) buying these boxes so they can be assured of a
staging area when they need it. The larger cluster is still shared but
once it comes to combing through the results or preparing for an intense
run, they like the convenience and capability of a local SMP box as
something of a "pre" cluster resource.

Steve