Beverly Clayton, who has served as the Pittsburgh Supercomputing Center's executive director since its founding in 1986, has announced her retirement effective March 31. She was one of the first people hired when PSC received funding from the National Science Foundation. During the early years of the center, she led the initial hiring of staff, coordinated the creation of office space in the Mellon Institute Building and developed organizational policies and procedures that set PSC off and running.
For many years, she directed PSC outreach efforts and has represented the center as a speaker at many public functions. Her activities have been instrumental in obtaining PSC funding from the Commonwealth of Pennsylvania that has totaled $30 million over 20 years. In 2006, she led the center's move into a renovated space at 300 South Craig Street and helped organize the center's 20-year birthday celebration. “All of us at PSC owe Beverly a tremendous debt of gratitude for her excellent, loyal service,” said PSC scientific directors Michael Levine and Ralph Roskies in an e-mail to staff.
HPCwire got the opportunity to ask Ms. Clayton to reflect on her time at PSC and to share her thoughts about managing a supercomputing center for over two decades.
HPCwire: Tell us a little bit about your background. How does one end up being the executive director for a supercomputing center?
Clayton: With a math undergraduate degree, I began my career as a Fortran programmer and Fortran teacher at Gulf Oil Research and Development. My interest in science was tangential to my interest in logic and math applications. At that time, there were very few scientists who could program, and even fewer women in programming. Gulf was a pioneer in hiring women in this area. Over the years as Gulf changed computer systems and made more use of them in both corporate and research areas, I changed assignments. Every few years I tried something new, including systems (also called assembly language) programmer, supervisor, computer education coordinator and finally, in the last days of Gulf, director and manager of datacenter functions. When Gulf left me, and went to California, I declined the transfer. I contacted several former colleagues in the Pittsburgh area, and one of them passed on my resume to Jim Kasdorf of Westinghouse who was one of the authors of the Pittsburgh Supercomputing Center's proposal which recently had been funded. He felt my background would fit with building a new Center and he recommended I be interviewed by Ralph Roskies and Mike Levine, the Center's Scientific Directors.
HPCwire: Besides the supercomputing centers themselves, high-end supercomputing is being done in DOE national labs, some top tier universities, and a handful of companies. What unique role does the supercomputing center serve?
Clayton: The National Science Foundation-funded centers support open, publishable research. The research proposed is peer reviewed and all disciplines are invited, and encouraged, to utilize the NSF centers. The PSC, for many years now, has invited and attracted those applications whose need is for large-scale computational problems whose solution cannot be done with any other systems, and which may not have even be considered solvable until now.
HPCwire: Looking back over the last 21 years as executive director of the Pittsburgh Supercomputing Center, how has the center changed over the years?
Clayton: Actually, I think it has not changed. It has grown, and replaced systems, and added new functions as they have become available, but the focus is still the same: Make available the most capable systems to the scientific community, and support reseachers as much as possible to enable them to make breakthrough achievements.
The only noticeable change is the growth of the staff and our new office location, which is wonderful.
HPCwire: Compared to the other centers, what do you think is unique about PSC today?
Clayton: The short answer is: the strength of our user support and our focus on the science.
For 21 years, PSC has been at the forefront in providing the most capable systems backed by outstanding support for the scientific community specifically focused on the very high-end users. The answer to the question “What is the measure of success of PSC?” is the science that has been done.
The vast majority of PSC's systems over the years were the first of their line available to the open community. Each new system required very substantial work by the PSC staff with the vendor to make it into a useful research tool, especially for high-end work. Each of them also required considerable PSC support-staff effort, working closely with the nation's top research groups, to produce important and practical scientific results from the new systems. Through this work, PSC developed the reputation as the center of choice for the most demanding users in the NSF program.
HPCwire: What's been the biggest challenge at PSC?
Clayton: The biggest challenge is the same as every other organization in a world of inconsistent funding. Like companies who need to continually innovate, develop and sell products or services to maintain market share, university and non-profit centers need to do this also. And we continually have to justify our existence, to a changing audience. Funding agency personnel, at all levels, change every few years, and the institutional memory is short. Political forces also come into play as leadership changes and the needs — real and perceived — of the state and country change.
HPCwire: If you could change one thing about how supercomputing centers are run or funded, what would it be?
Clayton: Provide longer-term funding. Centers should have frequent, thorough reviews to ensure funding is appropriately allocated, but ongoing support is really needed.
HPCwire: You've been the executive director at PSC for more than two decades. Do you think you'll miss it?
Clayton: I have a large list of “to-do” projects as well as a long list of places to visit in my Airstream travel trailer, but I do intend to maintain a Pittsburgh residence, so I will be dropping in on PSC to follow their continued success.
I'll miss the people, but I intend to stay in touch. I may do a little consulting and I will be working as a volunteer on the SC08 conference so that will enable me to stay connected with the community. This community of volunteers, those who developed the concept of the Supercomputing Conference series and have expanded and improved it over the years, is a fantastic group. While a few functions within the organizational tasks are now done by professionals, the core of the conference planning and the technical content and guidelines that drive the integrity of the conference remain the responsibility of the volunteers. I feel honored and proud to have been a part of this conference since the beginning, and have seen the growth of both the conference and the people working on it. The quality is still high, and the attendance keeps growing. The Education and Broadening Participation programs are valuable additions to the conference which are having an impact on the future of our profession.