[Biodevelopers] Re: [Bioclusters] Workstation Selection for Bioinformatics Research

[Joe Landman]

Hi Chip:

  The workstation choice is usually made after the software choice.  The
operating system choice is usually dictated by the software choice.  If
you have your software list, look at which sets of platforms enable the
software to run.  It might help to know what platform the software is
written on, as that invariably gets bugfixes faster than the other
platforms.

  You can buy a workstation (Dell, Sun, others) or put one together
yourself.  Questions you need to ask are

a) are you willing/able to support and service your own hardware?

b) are you running a "production" facility where machines becoming
unusable for any significant amount of time results in a loss of grant
money, or is at odds with a contract (this would be the case in a
service facility with service level agreements, not too many in academia
but I see them).

c) what level of performance do you need?  Do you have sample runs that
are representative of the workload?  Are you able to neatly package them
up and run them on such hardware?  

d) what configuration do you require to run your software?  Are your
models huge (lots of memory)?  Are you IO patterns well known and
optimizable?  Is your data local or remote?  Do you require graphical
display?

There are other questions that need to be answered as well, but if you
can answer these, you will be on your way to selecting the appropriate
hardware.

Joe 

-- 
Joseph Landman, Ph.D
Scalable Informatics LLC,
email: landman at scalableinformatics.com
web  : http://scalableinformatics.com
phone: +1 734 612 4615