Snow and ice fail to freeze battlin' Big Iron bots in their tracks
Four more student cluster teams face the El Reg cameras
SC12 It's time, once again, to meet some more plucky contenders in the SC12 Student Cluster Competition.
Team Slippery Rock
We meet the Slippery Rock team for the first time. They’re competing in the Mini Iron LittleFe division of the cluster-building contest. In the video they explain their strategy for achieving the best solution for the 10,000 city Traveling Salesman problem. Like most of the teams, they’ll be using multiple approaches simultaneously in order to determine which of the many algorithms available to them work best on the data set they receive.
This video was also recorded during the record-breaking snowfall and cold snap, meaning the temperature in the hall was approaching the freezing point in many locations. This explains a bit of the shaky camera work - incompetence explains the rest of it.
As the home team, there’s a lot of pressure on Utah to win. Well, maybe not – but I’m trying to create some for my own amusement. They agreed that there is some pressure, but the team members say that they’re resigned to living with the fear and using it motivate them to greater heights - or something like that.
In the video, we talk about how they’re going to attack the 10,000 city Traveling Salesman problem. Like other teams, they’ll be using multiple algorithms, but will be ready to change on the fly if one set of routines proves to be more successful. The team also has some of the best visual aids that explain both the problem and their strategy for tackling it. I’ll include those in an upcoming article once I get good digital copies.
To me, the Utes look like one of the stronger teams in the field – but what the hell do I know? In my real-life business travels, I’ve been stumped by seven and eight city trips – usually forgetting what my rental car looks like and being forced to wander through parking lots pressing the alarm button on the key fob until I spot my car.
Team Staten Island
Team Staten Island takes us through the LittleFe system and discusses their configuration and capabilities. These systems are small but they pack quite a punch when they’re teamed up to attack a single objective. These guys have their own LittleFe at their home base in New York and are obviously familiar with both the system and the Traveling Salesman problem they’ll be confronting in the challenge.
As explained, this problem is going to include 10,000 cities, meaning that a brute force approach would require a system to consider 7 to the power of 150 potential permutations. That’s enough to challenge even the largest of supercomputers.
In the accompanying video, I interview the first high-school team to ever compete in the Student Cluster Competition. They’re the team from hometown Skyline High School even though they look college age to me. They’re a bit quiet at first, but over time, I was able to draw them out a little bit.
We went on to discuss their unique approach to the Traveling Salesman problem. Rather than rely on the commonly available algorithms used to solve these type problems, they’ve decided to go their own way with a mostly homegrown solution. Moreover, they’re not going to be using multiple approaches at the same time - they’re just going to plough ahead with their own tool and modify it as needed along the way.
It’s a bold move that could either pay off in a big way, or leave them trailing behind the pack. They’re a confident bunch and work well together as a team – and that counts for a lot in these types of competitions. We’ll see what happens. And I still think these guys are all in their mid-20s, regardless of the driving licences and other documents they want to flash at me. ®
You can find more coverage of the student cluster competition as well as SC12 supercomputer news right here
Sponsored: RAID: End of an era?