Call 1-800-HLP-DESK - Episode 3

Show Notes

[00:00:00] Intro

Nicole: Hello, and welcome to the long tails of science podcast, where we interview women in high performance computing about science research, mentors, and career paths. I’m your host, Nicole Brewer. In this episode, we talk with Nancy Wilkins-Dier about her pathway into HPC. We discussed her lengthy career at the San Diego supercomputer center. Her involvement with Terra grid exceed and science gateways. And we also discussed the state of software related contributions in academia. Finally, we end with Nancy giving a little career advice to students and we catch up with her on her recent retirement.

[00:01:08] Introduction

Nicole: Nancy as a former associate director of the San Diego super computing center. And she has been there since 1993. And she’s now retired. She’s held a variety of management positions and her particular expertise is in the development of web interfaces to high-performance computing systems, data, collections, instruments, and other resources, fundamental to many of today’s research endeavors . Beginning with the NSF’s TeraGrid science gateways program in 2005, she has led many programs that support the democratization of access to high-end resources. Most recently, she has served as co-PI on both the NSF XSEDE program and the Science Gateways Community Institute and NSF scientific software innovation institute. She has held a number of leadership roles in NSF projects funded over hundreds of millions of dollars over her career. Nancy received her bachelor’s degree from Boston college in mathematics and philosophy and her master’s degree in aerospace engineering from San Diego state university. Welcome to the show, Nancy, and thank you so much for joining me.

Nancy: Thank you very much, Nicole. Thanks for having me.

Nicole: So I’d like to get started with just getting to know you a little better. What do you like to do in your free time?

Nancy: As most of my work colleagues know I do a lot of running even before I was retired. I was well known for that. I used to organize runs at a lot of the major conferences like supercomputing and PEARC, and basically anywhere I went to a conference, I was organizing some kind of run. So I got to meet a lot of people from a lot of different institutions and have some really great experiences out on the trails in salt lake city or through new Orleans bourbon street in the early morning hours. So it was a source of meeting, lots of great colleagues, some of whom I still keep in touch with. That’s pretty cool. I actually went to PEARC and I’m sure that you were the one that organized it in Pittsburgh, there was a run. Yup.

Nicole: That’s awesome.

[00:03:22] Pathway to Compting

Nicole: Let’s start with your initial experience in technology and computing in general, because I think it can be interesting to hear the perspective of women because the story doesn’t always start with, I was five years old and I built a computer.

Nancy: No, not at all.

Nicole: So what was your first exposure to computing and what got you interested.

Nancy: Yeah, it was kind of a non-traditional role, really. Like you said, I wasn’t building computers at five. I didn’t have this lifelong interest in science. When I graduated from high school about the only thing I knew about what I wanted to do is that I liked math and that came easy. It was fun. So that led to a college degree in math and it wasn’t until the end of the. Four years of undergraduate. When I actually realized that math could be applied to things, we proved a lot of theorems. There was not practical applications of math and the program that I was in. So we finally had this professor from bell labs senior year, and then I realized it could be applied to engineering. And that was like a revelation to me. I was very sheltered.

Then I decided to go for a graduate degree in engineering. And when I was first starting at San Diego state, the boyfriend of one of my college friends who was at Stanford, also in an aerospace program, and he was working with Robert McCormack, who is one of the early writers of computational fluid dynamics codes. This was in the eighties. So this stuff was just getting started then there wasn’t really a whole lot of computing going on back then, but I was fascinated by this and the fact that you could model physical processes with math and then programming it into something that computer could understand, visualize the results I thought was really cool. That was way before the great viz that we have today, we had some pretty simple stuff back then, and I was still impressed, but that’s how I got into it. I was fortunate to have professors who kind of let me do my own thing at San Diego state, cause nobody was really doing that. So I was able to pave my own way. I use the academic computing systems at night when they weren’t being used to process student records.

So I had this account on a VAX that I could run on all night long, but not during the day. And I had a few precious of Cray time at the San Diego Supercomputer Center that I would say for like really special occasions. I mean, I had like an hour time, total and a quarter or something, which is like, just like laughable now compared to what people had. But we had these block grant programs that went to universities, I think in the beginning it was throughout California. And now of course it’s everywhere and there’s the campus champions program and all these great things. But it was kind of funny because I was working out of a basement lab where I could have 24 hour access to a terminal and we couldn’t make phone calls from the phones there. You could only call within San Diego State could just dial extensions. And so anytime I had a question, I had to run up to the payphone in the courtyard at the engineering building, put it in my little quarter and ask a question. So one of the first things that I didwhen I finally got to work at the supercomputer center was installed an 800 number. So other people didn’t have to do what I did to get their questions answered. So it was kind of roundabout, but…

Nicole: I just finished Outliers actually and the book talks about how, when you think of all the tech people that started all these companies, a huge part of their success is always how many computing hours that they can get in on these machines back in the day where there’s so many physical limitations, like being in the right room and being able to make a phone call. So that’s super enlightening and very cool.

[00:07:12] Transition to San Diego Supercomputer Center

Nicole: So then from there, how did you start getting involved in this sort of intersection between software and engineering and science?

Nancy: Yeah. So my first couple of jobs out of grad school were at engineering firms here in San Diego. So I ran computational fluid dynamics. CFD is the acronym I’ll use cause it’s easier. CFD simulations for missile design at General Dynamics, which was a really large employer at the time I graduated. And then moved on to General Atomics where I was doing designs for gas, cooled nuclear reactors, so helium flow through a reactor instead of airflow over a missile, but similar kinds of concerns and codes and things like that. And both of those programs ended, there was like a shutdown in funding for the missile systems at GD and the same thing for the reactor program at GA. And at the time general Atomics was running the San Diego supercomputer center, which sits like across the street, on the campus of UCLA. So it was an internal transfer, just looking for another position and ended up getting the position at SDSC and then staying there for like 25 years.

So there was so fascinating and I liked it so much more than engineering, especially in a large company, you can get really pigeonholed where you’ll do the same thing over and over in slightly different ways as they make a little design changes. And with supercomputing, people were using the supercomputers in all, all areas of science, not just engineering, chemistry, biology, and now it’s even social science is just so thrilling. And so there I was never bored. I had a position answering the phone when I first started the help desk, the same number that I called as a student. So that was kinda cool, but just talking to people from all over the country, trying to do all sorts of different things and then having new technology and new computers to contend with really every couple of years, I mean, it’s easy to have a career like that.

Nicole: Well, it’s a trade off, right? It’s sometimes in the academic side, you know, we don’t get paid quite as much, but there’s people here for a reason. There’s something intrinsically interesting about being in this intersection and it’s really dynamic, I think.

Nancy: Yeah. I agree. Yeah. Yeah. Really rewarding too.

[00:09:34] TeraGrid

Nicole: So a lot of your work is focused on software that makes computing systems more accessible and maybe TeraGrid is the first installment of that.

Nancy: Yeah. I mean, you can’t always plan out a career and how it’s going to go and what twists and turns that are going to be there. And that science gateway program really changed my life, changed my career path. And Charlie Catlett, who I think is still at Argonne National Lab to thank for putting me in charge of something like that. So that was kind of early two thousands and the web was just getting going. And prior to that, people accessing super computers. You didn’t have to be at the facility of course, but it was very much interfacing at the command line and working very directly with the supercomputers. So we’re starting to see, the web develop and scientists using the web. You have these large communities of people who need this processing power, but don’t have all the background training in HPC or all the expertise. And so that was kind of the genisis of that program was looking at some of the large NSF projects that were web based and how we could work in a very hands-on way with eight projects to what did an interface to a supercomputer have to look like so that you could go in from the web and they, I want to do this simulation. I’m going to construct this Fortran code and I’m going to submit this code through a web interface to the computer, but even stepping back from that, like, I just want to solve this scientific problem. And describe it in terms that I know how to describe. And then on the back end, it’s going to construct the code and launch the code and then deliver me the result. And it was a very new way of accessing super computers.

We had a large working group, including security professionals. We had to make policy changes with the national science foundation for how these things can be accessed, because you can imagine there’s like a lot more anonymity and a lot more trust that has to be there to use these expensive resources, but we were able to navigate all those roadblocks and create a very successful program that still operates today. Now there’s hundreds and hundreds of gateways that use the supercomputers that the NSF funds. Yeah. I’m really proud of that, but it was all just sort of, you know, luck. Cause I think I was involved in user services still with like the consulting aspect of supercomputing and him just saying, “Hey, how’d you like to do this?”. Okay.

[00:12:03] Transitioning to XSEDE

Nicole: You’ve been involved with both TeraGrid and XSEDE. So maybe we step back and talk about what the goals of TeraGrid were and then maybe how that’s changed now that we’ve transitioned to XSEDE.

Nancy: So back then grid computing was more of a thing. Now it’s more cloud computing that we hear about, but the very initial idea was to construct a set of distributed machines that were identical, but geographically located apart from one another, and that people would run large jobs across all of these machines at assembled. So there was a great emphasis on the networking.

But really what TeraGrid did was bringing together these different supercomputer centers in non-competitive way. So working together on a single project and so XSEDE has carried that idea forward through John town’s leadership. And I think it’s been a huge win for the centers. Back when I first started at SDSC, we were very siloed. SDSC that Pittsburgh supercomputer center NCSA. Don’t talk to the enemy, man, no working with them. So TeraGrid was the first instance of a project. We were all funded to do the same thing together, and there were some steps forward and back, and I’m not going to say it was the smoothest transition from whatever years of working, one way to working another, but especially with John’s leadership through the XSEDE program and all the great people that work on it, we’ve made huge strides. Had lunch recently with Ralph Roskies, who is the director of Pittsburgh supercomputer center and his wife. And we get together regularly, even though we’re both retired now. So there’s been some huge wins through this programs.

[00:13:43] Science Gateways

Nicole: Wonderful. Yeah, so I’m told that you coined the term science gateways.

Nancy: You know, I wouldn’t say coined. We can’t remember who coined it. I would say I popularized.

Nicole: So for those people that don’t know what is the science gateway?

Nancy: I think of it as a web interface to some kind of capability, whether it’s data collections or. Analysis or remote instruments we’ve worked with telescopes and things like that. The gateways that I’ve always worked with have had that computing aspect. And so we coined the phrase in connection with computing, but the democratization, which I think was in the bio, that’s a word that we like to use a lot because there are an awful lot of groups that don’t have. Two large data stores and high-end computing and things like that. And if you can get access to a gateway and a lot of them are really pretty open and provide access to anyone anywhere in the world, you can get all these capabilities at your fingertips. It’s amazing. Some of the statistics from some of the gateways where they’ve just got worldwide users, some of them from not very rich countries, but still doing very important research. And they’re able to do that because of this gateway where otherwise, you know, no way would they have access to those sorts of things.

Nicole: Are there any particular gateways that you’ve worked on that you’re particularly proud of.

Nancy: The Cypress science gateway, which uses a lot of resources at the San Diego supercomputer center. That’s been a leader for a long time and they do phylogenetic tree analysis, which is useful in a lot of different fields. Nano hub is a very successful science gateway. It’s kind of interesting how some of these gateways have changed. A lot of things too. Cause some of them, it’s just individuals running jobs, running analysis, and not really interacting with other individuals, but some of the gateways really form these communities where people are answering questions for one another and contributing content like code or presentations or classroom materials, all kinds of different things.

[00:15:56] Giving Credit to Software-Related Contributions

Nancy: And in nano hub, for example, they’ve actually been able to change. The tenure process a little bit. So looking, not just at publications and citations, but looking at what someone has contributed to this gateway and how many of the thousands of people on this gateway have found those contributions useful and making that also a component. So I think it’s kind of modernizing a little bit, or what’s considered as a valuable contribution to science.

Nicole: Yeah. And there’s a lot of talk about how do we do that formally to make sure that people are sort of recognizing.

Nancy: Yep.

Nicole: Do you have any ideas on how that could be improved?

Nancy: We have tried to work with people like AltMetrics was a good group that I remember working with, but there were some people doing some really leading work, especially software contributions. Dan Katz had a lot to do with recognition of software contributions back when he was an director at the NSF, which were really important because you think of the value in creating software, that’s heavily used. And there was really no mechanism, especially for open source to kind of recognize how much something was getting used and how you were able to site software and things like that. So he’s done a lot of that of work in that regard. We’ve tried to do some work as far as citing the gateways because they have to secure their own funding. So they have to show their utility to their users. We were always trying to, to work with them in that regard to make sure that they were capturing success and telling those stories in the right way, because it was vital to their survival to.

[00:17:35] Project Managment

Nancy: I just wanted to talk more about your leadership positions and how management is also an important role. And we get divided up into people that want to work on tech and people that want to go into management and maybe why it was the choice for you. Yeah, I think I was good at it. I think it’s kind of interesting working with a lot of tech folks. A lot of people like to be more heads down programming instead of lots of communication with lots of people. And I think it’s great that we have so many different types of people like that because we couldn’t, you know, absolutely couldn’t achieve what we do if everyone were the same way. Right. So I think it’s a good fit for me. And it has been from early. It helps desk stuff for a few years at SDSC, but really within the first five years, I was starting to do management of different programs like back to impact E gosh, in the nineties. But project management, working with large teams. That’s been something that I’ve been good at, and it’s kind of been rewarding, especially in XSEDE’s ECSS program, where we had a large staff of people across close to 30 FTEs, maybe 70 individuals across. ten-ish sites across the country. And what we’re doing is matchmaking requests that come in for a supercomputing support. Could be gateways, could be optimization, could be like development of community codes, all sorts of different things or development of training and matchmaking, our staff to these projects. And then to be able to see them go off together and do great things. And, you know, you know, following up all the interviews at the end of these projects, how did things go? You know, is there anything we can improve and just kind of working on a program like that? It, something that, that I think I was, I was well-suited to, and it was very rewarding because you felt like you got to do more. By allowing all these other people to do more, if that makes sense.

Nicole: Yeah, it sounds like communication and looking at things from the high level are really important.

Nancy: There’s need for all of that in technology, there’s definitely need for the depth of knowledge in particular programming techniques and things like that. But there, there is also the need to step back a level a level, and be able to explain things to the community, explain things to review panels, you know, write compelling proposals, all those sorts of things. So a lot of different skills.

[00:20:14] Retirement

Nicole: What are you most proud of that you’ve worked on?

Nancy: I think changing the fundamental access to supercomputers the NSFs recognition that the gateways were such importance, that they funded a software Institute around gateways and their success and helping, you know, all kinds of different gateways to be successful. That I think I’m most proud of. That was my final, big award before I retired and just work with a tremendous group of people on that project. So having something that you’ve created that’ll live on without you, and sort of developing enough good people to just kind of carry that, what I consider important work, forward. Very satisfying, very fortunate to have been able to do that in a career.

Nicole: Absolutely. And congratulations on that, and congratulations on retiring.

Nancy: Yeah. Yeah. My husband predated me in retirement by some eight years. So he was anxiously after me to join him. And it’s been a lot of fun. I’ve been able to pursue a lot of personal interests. Like I, I took a math class at Stanford through Coursera because I hadn’t done math in a while and I missed that. You know, trying to improve my French, learn the oboe. Maybe I’ll go back to doing some programming. Python’s kind of on my list and never really got a chance to learn that. So. That stuff’s fun. That’s amazing. I mean, have you always just been working even when you’re done with work for the day? I mean, no, and that’s kind of the fun of retirement because I mean, I did work pretty hard when I was working on it, so I was pretty exhausted whenever I wasn’t working. So it’s nice to have the energy and the time to know, take off some of those enjoyable things, you know, now.

[00:21:49] Advise for Students

Nancy: Do you have any last thoughts or advice for the younger people listening? I gave a talk at my high school a while back, and one of the things that I said to those students was to not feel like you had to have everything figured out for your life, that it’s okay to kind of take it one step at a time.

I mean, as an undergraduate, I hated programming because you couldn’t wait till the last minute to do your homework. Cause it always required a little more time to get things, to compile and run. So I just, I didn’t like it. I, if you would have said, you’re going to be working at a supercomputer center for most of your entire career. I said, no, I’m not. So just be curious, be open to new ideas. Don’t I think there’s pressure now to feel like you have to have everything figured out. So don’t, bring that pressure on yourself and it’s okay to just like the next step. I’m going to explore math and next step, I’m going to explore engineering and for you, it might be totally different steps, but you know, don’t panic about that, I guess I would say. And technology is a great career for women. It’s lots of fun!

Nicole: Thank you so much for coming on the show. Nancy, I learned a lot about science gateways and I really appreciate your time. Thanks for coming on.

Nancy: Thank you very much, Nicole. I appreciate being asked.

Nicole: Thank you so much for listening to this episode of long tails of science. This podcast was produced by women in HPC at Purdue, an organization dedicated to promoting and advancing the representation of women in high performance computing. We are a chapter of an international organization by the same name, and you can sign up to be a member at womeninhpc.org. Follow us on Twitter or sign up for our email lists to keep up to date with the new podcast episodes and semi annual virtual meetings. If there’s a guest speaker, including yourself, you would like to nominate. Please send us an email and finally subscribe and rate the podcast on your favorite platform. Or listen on the web at www.breaker.audio/long-tales-of-science

Until next time, I’m your host Nicole Brewer. And it’s been a true pleasure introducing you to amazing women in science, engineering, and technology.

Guests

Hosts