Medical Area for Beowulf ?

[Schilling, Richard]

Typically in health care, we utilize minis and mainframes to handle medical
data - most of our computing is for business use at hospitals (medical
claims processing), rather than scientific - it's transaction based.  Far
easier at this point to just purchase and license (even at a cost of seven
figures per year) the big iron then spend time developing code.  

We get our software almost exclusively from vendors - so we're at their
mercy architecture wise.  For lighter work, like image processing, the
higher end NT boxes seem to be "good enough", which is what clinicians and
CEOs seem to be most interested in.

Our newest system which will be installed within 18 months utilizes about 30
nodes, but I'm not sure how the processing is handled.  I think it's an
embarrassingly parallel approach, rather than distributed algorithms.  In
general I'm not sure if anyone's done any research yet to determine if
health care transactions are best handled in a distributed or parallel
fashion.  We'll see.

I've been tinkering with a small 3-node Beowulf (for starters) to process
medical data and will eventually do some testing with patient benchmarking.
As our PCs get retired, I'm grabbing them and making use of them.  We should
be retiring about 24 of them here in a few weeks, so I'll have some more
fun!


Richard Schilling
Webmaster / Web Integration Programmer
Affiliated Health Services
Mount Vernon, WA USA
http://www.affiliatedhealth.org


> -----Original Message-----
> From: Kevin Rosenberg [mailto:kevin at rosenberg.net]
> Sent: Tuesday, February 13, 2001 11:06 AM
> To: Yoon Jae Ho
> Cc: beowulf at beowulf.org
> Subject: Re: Medical Area for Beowulf ?
> 
> 
> > I think, our beowulf computers can be applied to the medical area.
> > but I don't know the beowulf application examples for the sick.
> >
> > Anybody here had a experience or have a plan to apply our 
> beowulf computers to the medical area
> 
> I use MPI processing as an option in my open-source computed 
> tomography
> simulator: CTSim (http://www.ctsim.org). I've tested this 
> application in a
> 16-CPU beowulf cluster with good results.
> 
> I've heard from a physician in Germany that he is using CTSim 
> to remove
> artifacts for metallic objects prior to reconstruction. 
> However, I'm not
> sure he is using the Beowulf-enabled version of CTSim or the 
> single-CPU
> graphical user-interface version.
> 
> I look forward to hearing other responses to your query.
> 
> --
> Kevin Rosenberg, M.D.
> kevin at rosenberg.net
> 
> 
> 
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) 
> visit http://www.beowulf.org/mailman/listinfo/beowulf
>