[SATLUG] Next month's presentation

[John Ziriax]

Sadly, I have no Beowulf at home, that would not contribute to domestic bliss.

Work is another story.
There are several members or our group lurking on this list.
Maybe Vern would what to help with a Cluster presentation? What do you say 
Vern?


Our first one was a 24 node MPI based system consisting of 233MHz Pent. with 
256MB ram that were about to be surplused. Boy did that system surprise the 
powers that be. "You made this out of what?"

But, its gone now.

The second one is an 18 node 500MHz Celeron system with 1GB of memory /node. 

The third one is an 40 node system, recently expanded to 67 nodes (I think). 
These nodes have between 750 MB and 1 GB of memory per node. The CPUs are 750 
MHz Athlons. It has just been expanded so, I'm not too sure of the exact 
numbers. 

I'm currently building 9 more nodes of dual-Athlon MPs 2100+ with 3GB per 
node.  They will probably join the third system in some way.

What do we do with all this power?  And heat?  We simulate the absorbtion of 
electromagnetic waves (radio and radar frequencies) in a 1mm resolution model 
of the human (he's got over 365 million voxels). 

We are using LAM-MPI and Layey Fortran 95 Express.  And not much else. Since 
these are not really shared resources.

Since the output data files are around 1.5 GB, I'm always looking for data 
analysis tools.

It is nice to know there are other Beowulf hackers out there.



On Tuesday 27 August 2002 07:54 pm, you wrote:
> > Good idea.  How many SATLUGGERS use a cluster(s)?  Jeremy and I make 2.
>
> I thought I was the only one, cool! What kind/type and what purpose is
> yours? I actually have 2 clusters, one at home and one at work.
>
> Home (2 node): Dual P3-450, Debian 'Sid', 256megs RAM, my multimedia
> capture/encode/edit machine. General all purpose server, openMosix 1.5.8,
> PVM and MPI  --- node2 is a 1.2Ghz Dell laptop, openMosix 1.5.8
>
> Work (20 node): Dual P3-500, Antsd, Condor, PVM, MPI, PVM-Povray,
> Distributed John - used mainly for Monte Carlo simulations with
> ultracentrifuge data, X-ray cyrstalography simulations and whatever else I
> can think of ;)
>
> Soon to come is parallel BLAST as soon as I work out the bugs.

-- 
John
------------------------------------------------------------

They that can give up essential liberty
to obtain a little temporary safety
deserve neither liberty nor safety. 
   Ben Franklin, Historical Review of Pennsylvania, 1759