Darkly Colored Felines of Fury

Author

Patrick Aud­ley

Loca­tion

56.457478–2.987452Uni­ver­sity of DundeeDundeeScot­landGB-SCUKhttp://www.dundee.ac.uk

Updated

2002-05-30

Date

A Career in Bioin­for­mat­icsa foray into Acad­e­mia2002-04-0156.457478–2.987452Uni­ver­sity of DundeeDundeeScot­landGB-SCUKhttp://www.dundee.ac.uk

I started work­ing at the Uni­ver­sity of Dundee after I moved to the UK to be with Eliz­a­beth. It was, in ret­ro­spect, fated to be. I was spend­ing all my time in Elizabeth’s lab learn­ing every­thing I could about mol­e­c­u­lar biol­ogy (thanks Eliz­a­beth!) and invari­ably ended up bump­ing into the only other peo­ple on cam­pus that had an inter­est in large scale com­put­ing. After a very short and infor­mal inter­view I started in the Com­pu­ta­tional Biol­ogy Dept. with a larger role over­see­ing all the high-performance com­put­ing in the fac­ulty. This doesn’t sound ter­ri­bly inter­est­ing until you con­sider the research pow­er­house that the Life Sci­ences Fac­ulty con­tained — to me, it meant being involved in projects with a huge range of research and some incred­i­bly inter­est­ing prob­lems to solve.

This job was as large a change as mov­ing to the UK in it’s own way — the pace was ter­ri­bly dif­fer­ent from the cor­po­rate world that I’d just left behind. After a brief set­tling period it was truly a dream job. Look­ing back, I think I enjoyed Acad­e­mia most because I got to work with smart peo­ple who were always focused on results instead of just pass­ing the time (pub­lish or per­ish!). I had my own super­com­puter (see below) and lots of cool hard­ware (some of it very strange and won­der­ful) along with access to all the facil­i­ties at the fac­ulty. That last one alone was worth the job alone, which tells you some­thing about me.

For the geeks in the crowd, this marked the first time I had access to over 3TB of stor­age (even­tu­ally over 3PB, yes, that’s 3,000,000 GB) and a few 100GHz (even­tu­ally, much, much more) of com­put­ing power. The only thing that was a down­grade from my gear at GT was the net­work which, at it’s peak, reached only around 5 GB/s.

I do have the hon­our of being, as of leav­ing the job, the per­son with the fastest ever NFS imple­men­ta­tion that IBM’s HPC teams have seen in terms of actual wire speed file trans­fers — 2GB/s of actual, real file trans­fer from a sin­gle SAN (fronted with 2 paired p-series NFS servers) over NFS to a clus­ter of com­put­ing nodes was a new sus­tained load record. Err.. ok, that was pretty geeky, but I was in Acad­e­mia after all, geek­i­ness is king there.