Student Views Supercomputing in engineering 
Supercomputing in Engineering
By Chris Malton
Smallpeice, students, sunny skies, and Southampton. That was the start of a wonderful 4 days in Southampton University. Having been on previous Smallpeice Trust courses, I was ready to get stuck in to what I knew would be 4 days of action packed fun activities, this time on the theme of Supercomputing in Engineering. This appealed to me since I'm interested in computers, and also engineering, and this course seemed like a blend of both. After the usual registration process and leaving our luggage in the rooms, we made our way onto campus for the start of the course.
After being introduced to the academics in charge of us for the week, Dr Kenji Takeda and Prof. Simon Cox, we settled down to our first task, understanding what we were going to be doing. This was relatively straight forward, and without further ado we set to work on the all-important PowerPoint presentation to explain what we had done this week. Kenji was eager to show us the flight simulator set up in the design lab, where aircraft can be tested easily, and without expensive costs. Quite simply, it was amazing. I never knew that it was so simple to fly an aircraft over the waters of Southampton. Then I tried landing an A380 at Heathrow. The simple answer is don't try to. It's such a huge aircraft that the moment you try and tilt it, you go off course.
Tuesday was very a much a hands on day, with us frantically trying to build a supercomputer out of standard PCs. The net result, a machine where individual nodes (the PCs we build) got sent back because of problems, such as lights too close to memory, etc etc. Eventually we ended up with a supercomputer that worked. 24 cores and 24GB of memory, and a load of disk space later, we had our supercomputer. It was impressive, to say the least.
With a supercomputer ready to process data, and us eager for more action, it was time to go for supper. Supper was a lively affair, since it was a well known fact that there was to be an Xbox 360 gaming challenge following it. Needless to say, I was pretty useless at it, and resorted to watching others go wrong about as often as I did.
Wednesday was all about aircraft design. With an ever growing need for newer, faster, cleaner, more cost-effective airliners, the designs for these gradually get more and more complicated. With the help of software such as Fluent (a fluid/object simulation package), it is possible to predict what will happen with certain aerofoil designs, whether they will get enough lift, or cause too much drag, and so on. With the help of a fantastic website, www.futureflight.org, it was possible for us to design an aircraft very simply, but with enormous amounts of feedback as to how good it was in various categories. It does not take a genius to figure out how to use it either, and we were quickly designing many aircraft (often failing to meet the criteria). Once we had designed the most efficient/cheapest/quietest aircraft, it was then a case of building it in a software package which nobody knew how to use, and this meant a steep learning curve for everyone. Despite this, once people got the hang of things, it was relatively straight forward, apart from our group. Our group managed to get very unlucky over the 4 days, ending up with a cluster node that failed to report diagnostics initially, and then the design package kept crashing. We have no idea why, it just did!
While all this was going on, it was a case of finishing the PowerPoint slide show for the course dinner display. This was straight forward enough, except I wanted to include a picture of the aircraft we had supposedly built in the design package on the last slide. This proved interesting given that the software kept crashing before we saved! Eventually we managed to wrestle the software onto another computer, at which point things started to look up. Of course, designing the aircraft for the umpteenth time was a little tedious, but all good practice. After a successful CAD design stage, we transferred it back to a PC with Microsoft Flight Simulator to export it to Flight Simulator. And then we got the most useless error imaginable: “Error exporting .MDL file”, followed by several more errors. Result: No flight simulator model. What did they say about bad things coming in threes?
This was all followed by the course dinner and a disco. At least, it would have been a disco if a disco could be defined as rocking at Guitar Hero. The course dinner is always interesting, and it was nice to see some faces I recognised from last year's Biomedical Engineering course I did with The Smallpeice Trust. It was also an ideal opportunity to talk with people in industry, such as Dr Newberry from Microsoft's High Performance Computing division.
Thursday morning was spent frantically rebuilding the model for the flight simulator, which fortunately was managed, and despite all our fears, flew flawlessly. At least, as far as we could tell it was not going to fall apart.
We then presented our short presentation to the Dean of the Faculty, and also Dr Michael Newberry, from Microsoft. While they were impressed with our presentation, group 5 were slightly ahead of us, and won the course prize. Since we were “doing stuff to do with aircraft”, Kenji figured he'd buy remote controlled model helicopters for the winning team. I dare say I'm envious!
After much hard work, and also fun, it came to the end of the course. The impression I got as I left was that everyone had a great time and didn't want it to end. Certainly I was not quite ready to leave, but that was only because I'd had such a great time and met some very interesting people. A number of us are still in touch with each other thanks to Facebook.
This unfortunately marks the end of my time participating in Smallpiece Trust courses. Over the past three years I have taken part in three courses run by the Smallpeice Trust and every one of them has been fantastically well run, and it's been great fun. Even if you have not got a huge interest in the subject, as was the case when I went on the biomedical engineering course in the summer of 2007, everyone is really friendly, and you'll enjoy most, if not all of the course.
