“How do you explain an HPC service in thirty seconds?”
At CIUK 2025, this was the final test for teams taking part in the part-one finale, hosted by Alces Flight, for this year’s Student Cluster Competition. After building, operating and supporting a real HPC environment over the course of less than three hours – and not to mention the weeks spent prior learning and qualifying for this moment – students now had to distil everything they had learned into a single, 30-second pitch.
The result is precisely what you would expect under pressure: fast thinking, creative explanations, nervous laughter and a fair bit of silliness. That’s all part of the fun. But behind the humour is something awe-inspiring.
Each team built a functioning HPC service from the ground up. They powered on real hardware, deployed an operating system, configured scheduling, installed applications, supported their users, and adapted their platforms to improve overall use. The thirty-second pitch was the culmination of their efforts – a bit of marketing wrapped in some understanding of what they had built, how it worked, and who it was for.
This supercut captures the joy and the messiness that make up a time-bound HPC competition. But under the hood, the students experienced the authentic flavour of what it is like to operate HPC services in production.
If you are interested in the technical and operational aspects of the challenge, we’ve published a short technical white paper outlining the architecture, design decisions, and open-source methodologies used to deliver the service-focused HPC environments at CIUK 2025.
Watch the pitches below, and download the technical white paper to explore how it all came together.