[Bioclusters] Clusters for bioinformatics... Some numbers

Jeremy Mann jeremy@bioc09.v19.uthscsa.edu
Thu, 30 Aug 2001 22:09:17 -0500 (CDT)

> Part II ... more questions ...
>> My problems with the NOW approach have little to do with power and
>> more  to do with (a) unpredictable available CPU power 
> What do you mean by that?
> Do you mean that the foreground user may steel all the CPU power from 
> the background process once in a while?  Does that have any negative 
> effect on the background job (other than steeling some time) if the 
> backround code is not parallel?

Not if you properly setup the cluster. If the cluster is YOURS on YOUR
network on YOUR switch, you have nothing to worry about. Just make sure its
on its own switch so you don't drown your own network traffic with cluster

> In which sense is the administration of a network of, say, 20 machines 
> harder than the administration of a cluster of, say, 20 nodes?

Very little. Our cluster has been running for about 2 years. The only
problem I come up with is replacing power supplies and power supply fans.

> Just to give you an idea of how inexperienced an naive I am: I thought 
> setting up a beowulf cluster with all the special cluster software and 
> Mosix on top and whatever else would be some extra work, making the 
> total installation and administration work *harder* and not *easier*?

Well once you get all the software and nodes talking to each other, its
pretty much gravy after that. You just need to find software to run on. You
might look at a package called OSCAR. Its all the packages you need (you
need to preinstall Redhat 7.1) to make your own cluster. Comes with a nice
looking GUI app. All you do is follow the instructions and it does
everything else.

>> network performance can do much more than slow a system down; it can 
>> cause jobs & data to disappear and other nastyness.
> Could you please give a few more details on those nasty things, and how
>  frequently they may happen?

That is precisely why you use high speed interconnects between the nodes

Jeremy Mann