Faces of HPC: Andrew Turner
Andy was the first in his family to attend university and now works at EPCC at the University of Edinburgh as a Project Manager. He is involved in various projects with HPC facilities, including ARCHER, DiRAC and PRACE. His work includes managing teams, writing scripts, and offering user support.
Andy came to HPC though computational science, with a bachelors in chemistry from the University of Wales, Bangor. He earned his PhD in astrophysical chemistry from the University of Liverpool, and has held research positions at the University of Edinburgh and the University of Strathclyde. Finding a personal preference for computation over experimentation, Andy shifted to HPC, and is currently a project manager at EPCC. His current projects include work with ARCHER, PRACE, and DiRac.
Tell a bit about yourself.
My name’s Andy turner, and I’m from Liverpool originally. I came to HPC through Science - originally computational chemistry. In school I thought chemistry was great, and I would become a chemical engineer. But I actually found that I was rubbish at the experiments! So by the time I was at university and doing chemistry, there was a computational final-year project on offer and I went for that. I really enjoyed learning how to program, so I went on to do a Masters and a PhD in computational chemistry.
I then did a post-doc. Although I was interested in the research, I really wasn’t interested in writing proposals, and it was just a pain and a drag applying for funding. I found that I really enjoyed the computational aspect.
Outside HPC, I mainly go cycling and mountain biking. When I get the chance I go caving underground. The best caves are in Northern England – in the UK, anyway.
What is your current job? Describe what you do in HPC.
My job at EPCC is Project Manager. I have a good mix of doing technical stuff – programming and scripting– and also helping with projects and managing teams of people.
One of the main things I’m involved in at EPCC at the moment is ARCHER, the UK national super computer. I head-up the computational science and engineering team. If somebody comes to ARCHER and says “my code needs porting or improving the performance,” we’re the team of people who help with that. So I do a bit of support myself, but I also manage those teams as well.
I’m also involved with the other national super computer in the UK, which is the DiRAC Supercomputing Facility, used by astrophysicists and particle physicists. You can’t really understand the results that come out of the LHC unless you have simulations, so we run that for them.
How did you become interested in HPC?
I liked doing the computation – for one, because it was very reproducible. In science, experiments are supposed to be reproducible, and that only works if you’re actually good at experiments– which I’m not! Whereas with computation, you just run the calculation twice. As long as it doesn’t have error or a random seed in it, getting the same result twice is just fine. It makes things much more understandable and controllable. You can very easily pick and choose your variables.
I also liked that in computational science, you could study things that you just can’t experimentally. In computational chemistry, you can study the chemistry in the centre of a star by observing it indirectly through the spectra that comes off a star. But HPC allows you to actually get down to the atomic level and find out what’s happening.
Is there something about you that’s given you a unique or creative approach to what you do?
One thing that I seem to be quite good at is getting on with people quite easily, although I’m actually quite a shy person!
Part of my background and my upbringing helps me. No one in my family went to university, and I was in a ‘lower achieving school,’ so the people there didn’t really go to university either. Because of this background, I’ve seen people from many different walks of life. I’ve interacted with people who aren’t university or academic-types. So I’m able to speak to different sorts of people and understand what they need.
I seem to be quite good at the user support because I’m good at understanding what the users need, and at helping them express themselves to find the proper requirements for talking about a piece of software that they can then use to scope out a project and take it forward. Just being able to get on with people and talk to them and make them feel at ease is the thing that helps me the most in my job.
Were there any challenges when you first entered the field?
The only thing I have sometimes had an issue with, working in science and computational science at the start, was because of my accent, though it’s not as strong as it used to be. Certainly at the start, it was getting people to take you more seriously. But I’ve since gotten older and more experienced, and people take you more seriously and give you more confidence. I guess I was less confident at the start, so perhaps it was an interplay of those things.
As far as challenges, I think it’s a lack of training. It’s a good thing as well! But, I’m entirely self-taught at HPC. I have no formal qualifications in computing, other than the fact that I did computational science for a while.
When I speak to people in this field, many have worked in HPC for ages and are trained in computer science. You’d be saying stuff to them and thinking, “I don’t even know if what I’m saying is true!” There’s that challenge of not having the formal training. I knew some elements really well and could talk about them in depth, but there were also gaps in my knowledge which wouldn’t be there if I’d had the formal training in the subject.
But that’s a strength as well! Because it means that there are always new and interesting things to learn on a day-to day basis.
What’s the best thing about working in HPC?
The best thing about working at EPCC is the opportunity to meet really clever and interesting people who do amazing research. But for HPC in general, it’s just the variety and the chance to learn new stuff all the time. That’s what I really love – that’s what drives me.
One of the things I enjoy best about my job is the diversity. You’ll be helping computational chemists one day, and the next it’ll be climate scientists, then it’ll be QCD researchers, and you get an overview of all these different areas of research, and how the computational techniques they use overlap in ways they aren’t even aware of. So you get somebody doing molecular simulation and n-body simulation, and they use exactly the same algorithms and techniques, at some level, as someone simulating cosmological simulations. I’m able to recognise that one technique for molecular dynamics could actually be really useful in cosmological simulations to make them run better. Having that overview and making them work together - It’s great, it’s very cool.
If there’s one thing about HPC you could change, what would it be?
I think it would be around funding in terms of national facilities. In the UK you have to reapply for funding for things like the national supercomputer for five years or so. And there’s always the chance you’re going to lose it! There’s no guarantee that EPCC will be running the next national supercomputer - we have to bid for it. And that makes it much more difficult than it should be to train staff, retain staff, have a coherent effort going forward, and a coherent body of expertise that can serve the scientists and the researchers who need that support to make the code work.
And there being no career path for people. For those who are interested in HPC, like me, you have to make your own career path. Almost anybody you talk to at EPCC here started out in something that’s mainly not HPC at all. They ended up there because they found they enjoyed it. Generally, without the formal training and the formal career path, it can be a challenge to get people into HPC and for them know the right things. That’s one of the reasons why we’re trying to develop these online courses!
What’s next for you in HPC – where does your career lead you?
Luckily here, there is a way to progress in a career. For me, I’m enjoying my role in ARCHER as a team leader and project manager the most. I’m now talking to people like the research councils about the future of EPCC in the UK and where it’s going. And I’m really enjoying the discussion of what’s important to the strategic idea of designing the next supercomputer servers, and what the UK needs in terms of skills.
I’d like to do more of that strategic work, making a success of HPC. I still want to be able to keep doing the technical work and do a bit of programming on the side – I don’t want to completely become a manager! But next for me is getting more involved with that strategic planning.
Andy was interviewed by Vivian Uhlir in August 2015.