Faces of HPC: Gavin Pringle
Gavin works at EPCC at the University of Edinburgh as an Applications Consultant. Some of his current projects include ARCHER’s Computational Science and Engineering support team, managing the FORTISSIMO User Support Helpdesk, and working with the Software Sustainability Institute.
Gavin earned his BSc from Napier University in Mathematics with Engineering Technology, as well as his PhD. He later did a post-doc at Napier in conjunction with Edinburgh University and the University of California at Berkeley. He currently works at EPCC providing support for various projects, including ARCHER, which includes working with the British Antarctic Survey simulating the cryosphere, FORTISSIMO, and the Software Sustainability Institute.
Tell us a bit about yourself.
I’m from inverness, but I’ve been in Edinburgh for 29 years. I was an undergraduate at Napier - it was Mathematics, Engineering, Statistics, and computer programming. My PhD was in fluid dynamics, looking at turbulent flow. My post-doc at Napier was four years, looking at the nature of turbulence - numerical turbulence - from a simulation point of view.
Outside of work I am currently attending the Edinburgh Fringe as my main hobby is theatre. In August, it is crazy as there are 3200 shows and I see about a hundred shows in the month, and I’m on number 72 at the moment.
What is your current job? Describe what you do in HPC.
I am an Applications Consultant here at EPCC at the University of Edinburgh, and I help academics and industry exploit large scale computing services not only in the UK, but across the world. I’m kind of on the client facing side of it, but I’m outward facing and inward facing as well - I do a lot of tech stuff, like taking codes and porting them to the National Supercomputers, and then optimizing them and making them available for UK or European researchers.
I am also currently working with FORTISSIMO. We work with industry to help them access HPC systems, but Industry is sometimes reluctant to use HPC due to perceived difficulties: they think they are difficult to program and difficult to use, and if they train somebody to use them, they worry that that person has become highly employable and they might get poached by another company. So Industry fears they'll be paying somebody to continually learn to use new machines. Now, there are lots of disparate HPC centres around Europe, all trying to promote HPC to Industy, so this FORTISSIMO programme is an attempt to centralise the effort. Its a central place where people from Industry can buy cycles, rather than having to go to each individual centre. What Im doing for that is ensuring the helpdesk is operated according to the agreed procedure. Its kind of client facing, but at the moment not actually addressing the clients.
HPC in general was something I kind of fell into. My main interest was theatre - I was directing and acting and I was in a professional run at the Lyceum theatre. People said that 70% of actors, no matter how good they are, are out of work at any given time, so every actor should have something to fall back on. So I fell back on my chosen safety-net career: I did applied mathematics for my PhD. Mathematics is not as far from theatre as it appears. You know that old idea of left-brain/right-brain? That’s been disproven recently, but it’s still a useful structure to hang a conversation on. Mathematics straddles both sides of the brain - you need to be disciplined but you also need the creativity. Same with music as well. I play Bass and Piano and Clarinet. You’ll find a lot of mathematicians are musicians and vice versa. I try to apply my maths wherever I can in my job.
How did you become interested in HPC?
Truthfully? I needed to fill 3 years to qualify for funding for a postdoctoral Directorial course at Bristol Old Vic, and I heard that one of the lecturers at Napier, Dave Summers, was going to California. I asked if I could carry his bags and he said, ‘Sure, you could become my PhD student’, and I said yes before I knew what it was about. Turns out, it was taking a new algorithm to solve the N-body problem using HPC. And that was before MPI existed, so there were only blocking sends and receives – really basic. I had to write my own collective communications routines, which are freely available now!
As part of this project we want to celebrate the diversity of HPC, in particular to promote equality across the nine “protected” characteristics of the UK Equality Act, which are replicated in world-wide equality legislation. Do you feel an affiliation with this matter, and if so how has this interacted with or impacted your job in the HPC community?’
I don’t come out to people as part of my everyday job, because indeed it’s none of their business, why would I? But I have been with people over a certain age who use terms which I find offensive because I’m gay. There are people quite high up in academic and industrial circles who come from a particular time when certain words or attitudes were commonplace, which are now unacceptable, thanks in part, to European legislation. Of course, discrimination was just rampant back in the 70s and 80s.
Because there are clients and we want their custom, I’ve kind of had to nod or smile or feign agreement when they offend. I don’t think it’s right for me to start talking about sexual politics in a field where it’s just not important or relevant. So you keep it professional, and you let it go, and you change the subject. You ignore it and hope it’ll go away over time. There’s a time and place to steer them, but business meetings aren’t suitable.
You know, a client took a phone call during a meeting and upon hanging up, he shouted ‘f***ing faggot’ at the phone. That was wholly unacceptable, and yet you just ignore it!
So is that really an impediment? No, it’s more of a life lesson to people out there… what did I learn from this? I learned there’s a damn good reason why the European regulations come in. I only fall under one of the protected characteristics, but other people must see that as well. You can’t stand up for your rights in a situation which could lose custom for your company. You’ve kind of got to kowtow to it. You’ve just got to keep going and ignore it. And in another time and place, somebody somewhere will hopefully educate that person.
People say the reason why they use such terms is because they come from a certain time. They’re not being vindictive or horrible, or they don’t know they’re being vindictive and horrible, it’s just the terminology they’ve always used. A lot of times, you don’t know what you say is offensive – you have to be told. But you don’t tell them when you’re trying to be helpful and work with them.
I wasn’t instructed to keep quiet, it was just something I realised was necessary. And I don’t believe it’s my place… that’s a difficult one. Because the university is bound by these EU regulations, and because I’m staff, we’re encouraged to promote the university’s philosophies. So outside of work, if you hear somebody breaking one of the protected characteristics, what am I supposed to do about it? You go up to some drunken guy and you end up in a fight? No, it’s clearly tempered by the situation.
So in that situation with the phone call I think I was right and I still wouldn’t say anything about it happening. I might say something like, ‘I don’t think that’s right, but anyway, let’s keep going’. I think it’s a gradual process and I don’t think people should shoot themselves in the foot in order to gain a fairer society. It just takes time. You can’t expect it to change overnight.
But this wasn’t a barrier! It wasn’t really a challenge, it was just sort of a raw introduction to how inconsiderate some people can be. But when you’re there, talking with a client, you’re not there to judge their personality.
It's not something I've overcome, per se, it's just something I've learned to tolerate. But tolerance is wrong! Once you tolerate a thing, then it means that that thing is 'bad' but you'll look the other way anyway. That sounds like inactivity. You've got to be proactive.
So some people say that ‘homophobic terms are used by a certain generation’, and I say ‘that’s the reason but not an excuse!’
Is there something about you that’s given you a unique or creative approach to what you do?
There aren’t many theatre people involved in supercomputing. I think that’s beneficial in that I can immediately connect with people. Growing up in a theatre you always worked with people of all ages, of all abilities, doing different jobs, and from different walks of life. I’ve never been ageist or sexist, for instance. It was just a whole bunch of interesting people, and I carry that around with me. I think everybody’s interesting. I think my theatre background helps when meeting clients for the first time. I can empathise pretty much straight away, and I can listen to their issues and be helpful, and make them feel like they’re being looked after and supported. Which is kind of what my job is, is to look after people and help them achieve their scientific goals. So I think that’s beneficial.
Were there any challenges when you first entered the field? How have you overcome these, or do they continue challenge you?
No, didn’t have any challenges. That is pretty positive!
What’s the best thing about working in HPC?
Meeting so many people from such diverse scientific backgrounds. And getting involved in projects I could never have even guessed I’d be getting into. Like intelligent buildings. So if there’s a fire, then the building will open or close windows or shut doors, assuming there’s no people trying to get out that way.
I think personalized medicine is also really fascinating. Everybody reacts to drugs differently: I read that Aspirin is a placebo for 25% of white people, and 80% of black people. Or some people have a small glass of wine and get really drunk and others can drink a bottle and not feel the effects. Also, everybody has a unique form of whatever virus or disease they have got. The unique mutation of the HIV virus is a motivating factor in the European effort towards personalized medicine as no two people have the same strain. So how do you fight such a virus when everybody has their own strain? And when you do find a drug that works for a whole bunch of people, it’s not going to work for everyone else. So in the future, we will provide a DNA sample in the doctors waiting room, and they’ll simulate the body and then give the virtual body a virtual disease and then try different drugs on that virtual you, and then they’ll find one that works for you and you alone. Then it’s manufactured in the doctor’s office! Then out comes the pill for you alone, and for your version of the disease. Isn’t that fantastic? And that’s 40 years away. I’ve contributed some work on that and it’s really interesting.
So yes, it’s coming into work and not really knowing what’s going to happen. It has a negative side as well – I’ve got a report that I’ve got to finish today and I’m here chatting with you! You never know what’s going to happen on the day. And it’s always going to be interesting and diverse, you meet a wide range of people and the vast majority of them are very nice.
If there’s one thing about HPC you could change, what would it be?
I always had the sensation that the vendors were trying to force the users onto machines that they don’t need. There’s a race to be on the Top 500 fastest machines. They build machines that will get them to the top of that list, but aren’t particularly useful or conducive to running the big or popular simulations. So the vendors are always saying, ‘here’s the hardware, if your software doesn’t run properly on it then you’ll have to change it.’ And HPC’s a loss leader for these companies. I think the only reason they’re making big machines is so they can have a big advertising name and sell lots of the cheaper equipment.
So I can understand their point of view from a financial and an industrial perspective, but from a scientific point of view, it’s not fair or right for the users. And a lot of the users are scientists, not computer scientists! What do they know about machines? So that’s kind of where I come in, to help them. I suppose I’m doing myself out of the job by suggesting this, but I think the direction of future architectures shouldn’t be driven solely by the demand or the desire to be the fastest machine. Or at least in the Top 500. It should be application driven as well.
That’s how it feels. Certainly a couple of years ago it felt like that. I think it’s getting better now, but I’m no longer sure.
What’s next for you in HPC – where does your career lead you?
Because the job is so diverse you never really know what’s going to happen, I can only say what I hope is going to happen. And what I hope is it’ll continue going as it is now. Being an Applications Consultant is so inclusive, you can manage small sub-projects, you can speak at conferences, and you can do teaching at remote classrooms… It’s so all-encompassing!
But I would like to do more travelling around Europe. I enjoy travelling as well, and meeting new people, and so I’m very happy to be involved in the FORTISSIMO European project.
Gavin was interviewed by Vivian Uhlir in August 2015.