• Question: when did you start working with super computers and did it take a long time to learn ?

    Asked by heft532shod93 to zongyuanzhu, richardfazackerley, Luke, craigmcneile on 19 Mar 2026.
    • Photo: Luke Humphrey

      Luke Humphrey answered on 19 Mar 2026:


      I first used a High Performance Computing (HPC) cluster (also known as a supercomputer) in 2020 during my graduate role at UK Atomic Energy Authority. We have access to a machine hosted at Cambridge University, but we can use it remotely remotely from anywhere by connecting from any laptop terminal.

      The interface is entirely command line using a Linux-based operating system, meaning there is no graphical user input. Instead, you need to type commands to navigate such as “cd documents” to “change directory (cd)” to the “documents” folder. Usually, I work by setting up simulation workflows locally in lower fidelity (this is similar to playing a video game on the lowest graphics settings so that it runs faster). This way, it’s a lot easier to find and solve bugs in an interface I’m more comfortable with. Then, when it’s ready I will port it onto the HPC cluster and submit the job.

      Because it’s a shared resource, there is a queue to run jobs, and every cluster might have some unique systems you need to learn based on the team that manages it. Their job is to make sure nobody is abusing the system or using up all of the computer at once without a good reason.

      I would say it’s not particularly hard to learn once you’re at the point when you’re comfortable on a command line. For a software engineer or computational scientist career, you will likely already be comfortable with that sort of interface by the time you get access to a supercomputer anyway.

Comments