Dr. Stephan Priebe, GE Aerospace Research
In our continuing series featuring promising young members of the HPC-AI community, we here feature Dr. Stephan Priebe, a Senior Engineer in the Aerodynamics and Computational Fluid Dynamics Lab at GE Aerospace Research in Niskayuna, NY. He became involved with HPC and AI in 2007 as a PhD candidate at Princeton, where he ran high-fidelity simulations of supersonic flows to study complex physics, including shock waves and turbulence.
This led to a long-term interest in computational fluid dynamics (large eddy simulation and direct numerical simulation) and the potential to leverage machine learning in turbulence modeling.
At GE Aerospace, he leads high-fidelity simulation efforts to study and model complex flow physics in jet engine applications, such as shock/boundary layer interactions in transonic compressors and 3D flow separation and vortex shedding in open fans.
Stephan has been PI and co-PI for multiple supercomputing allocation grants at the national labs since 2018. This includes recent work leveraging the Frontier exascale supercomputer at Oak Ridge National Lab to perform computational fluid dynamics simulations at unprecedented scale.
An Interview with Stephan Priebe: Exascale-Class Aerodynamics
What is your passion related to your career path?
Aerodynamics and turbulence are my main domain interests. Turbulence is a fascinating problem, one without a closed-form solution. It is a killer app for HPC and exascale computing, it allows us to simulate realistic flow conditions that involve a large range of spatial and temporal scales. As the world decarbonizes, the path to sustainable aviation involves improvements in the aerodynamic efficiency of aircraft and engines. The fundamental challenge is to accurately predict the behavior of complex turbulent flows in jet engines to enable those efficiency improvements. All of this would not be possible without HPC, and exascale supercomputing has enabled us to look at problems at full flight-scale (the previous generation of supercomputers could only do smaller problems at reduced scale).
Do you prefer working as an individual contributor or a team leader?
I have worked on some unique HPC simulations, including recent large-scale simulations of a novel open fan engine using Frontier, the world’s first exascale machine. These would not have been possible without a skilled team of code developers, turbulence and aerodynamics domain experts and hardware and visualization experts, as well as close collaboration with vendors. While I enjoy working as an individual contributor, I believe that true advances in HPC require a closely integrated team of specialists working together, as exemplified by our recent simulations on Frontier, showing the power of exascale. I have enjoyed helping grow the next set of HPC users, working with multiple collaborators and helping strategize how to grow the impact of HPC both in terms of productivity and capability. In addition to technical publications, I have also contributed to descriptions of our work to inform and grow the broader community.
Share with us work you’ve been involved with that brought about an advance, a new insight, an innovation, a step forward in computer science or scientific research.
I was involved in first-of-a-kind aerodynamics simulations on Frontier. Using up to 85 percent of the entire system, we performed highly resolved simulations of turbulent flow around an open fan for the next generation of sustainable aircraft engines. What is unique and impactful is that these simulations are at full scale conditions as encountered by an engine in flight and are therefore computationally significantly more extensive than the problems at reduced scale computed on the previous generation of supercomputers.
Who or what has influenced you the most to help you advance your career path in this advanced computing community?
I have received lots of support along the way, for which I am very grateful. As I described above, our work was done by a broad and skilled team. What has pushed me on is that advanced computing and high-fidelity simulations offer a unique and detailed view of turbulence, one that is required as we improve the aerodynamics of engines to address the challenge of sustainability in aviation.
Your thoughts on how we, the nation, build a stronger and deeper pipeline of talented and passionate HPC and AI professionals?
I believe it starts with those of us working in HPC and AI. If we can share our passion through outreach activities, communicate the science, fascination, impact and relevance of our work we will motivate the next generation to enter this field.
What does it take to be an effective leader in HPC and AI?
Perseverance – we are working on very challenging problems. Our simulations on Frontier were run on up to 85 percent of the entire system. These were some of the largest computational fluid dynamics simulations ever run, only possible on the first exascale machine. There were invariably lots of challenges the team had to work through and it took persistence to drive the project to success and demonstrate its impact. Identifying gaps and working together both with the internal team and external collaborators is key. Being aware of the rapidly changing HPC landscape and continuously adapting to the changes is also important to get ready for future developments and what comes in a few years.
What is the biggest challenge you face in your current role?
The biggest challenge is to extract physics insight from the rich and highly-resolved simulation data sets that we are gathering through HPC simulations. To address this challenge, we work closely with teams at Oak Ridge, for example, to improve our analysis and visualization, including leveraging AI techniques to develop lower-order models based on the high-fidelity simulations. Keeping up with the rapid evolution of hardware is also a challenge, as is how best to integrate AI and show its impact.
What changes do you see for the HPC-AI community in the next five-10 years, and how do you see your own skills evolving during this time frame?
I have no experience yet with quantum computing but it is something I would like to explore to see how it can help us with our problems.
Do you believe science drives technology or technology drives science?
Both, through new HPC systems and exascale we have powerful technology at our disposal that allows us to investigate scientific questions, such as turbulence, at a level of detail that was not previously possible. And scientific advances, for example in AI, are enabling the development of reduced-order models leading to new technology developments.
Would you like to share anything about your personal life?
I enjoy cycling and running and spending time with my family – my wife Pamela and my seven-year-old son Lukas.
Speak Your Mind