The Paths to HPC series, presented in collaboration with Women in HPC, showcases the women working in high-performance computing. Our hope is that by highlighting these trailblazers—and the sometimes unique paths they followed into the field—other women will feel inspired to envision themselves in similar roles.
What was your path to working with HPC?
As I was doing my PhD research, which focused on solving computing problems that arise in planning and operating energy systems, I realized that the problems I was solving required more computing power than my laptop had.
I was researching how to optimally size the capacity of a completely green microgrid, which is a localized group of renewable energy sources (solar panels and wind turbines) and loads that operate isolated or connected to the main grid. But because renewable energy sources produce different amounts of energy depending on the season of the year and the time of day, the problem had to be solved for many scenarios.
For example, solar panels produce more energy during the summer than in the winter. Solving this green microgrid capacity sizing problem required me to divide the problem into many sub-problems, and then solve the sub-problems in parallel on all the desktop computers I had in my lab. This was my first experience with large-scale parallel computing.
When I finished my PhD, I applied to a postdoc position at Lawrence Berkeley National Lab (LBNL) that focused on applying parallel computing in transportation simulation and optimization problems. Given my previous experience with solving large problems in energy systems, I thought it would be a good opportunity to apply my computing expertise in a different domain application.
I was fully introduced to HPC at LBNL, first by my postdoc advisor Dr. Sherry Li, and by taking training classes on various HPC topics. I quickly grew to enjoy working with HPC, as I gained experience in parallel computing and started running jobs on National Energy Research Scientific Computing Center (NERSC) supercomputers.
The coolest thing about working in HPC is that I get to use some of the fastest and most powerful computers in the world.
As a postdoc, I developed high-performance computing algorithms to solve large-scale traffic assignment problems. The results of this work contributed to starting the HPC4Mobility program at the Vehicle Technology Office of the US Department of Energy. My postdoc work also involved parallelizing a macroscopic traffic simulation on HPC systems, which resulted in speeding up the simulation by 198 times compared to serial simulations.
What’s cool about working with HPC?
The coolest thing about working in HPC is that I get to use some of the fastest and most powerful computers in the world. These supercomputers provide the power and memory for fast simulations and solutions and enable me to tackle large-scale problems that would otherwise take years or months to solve on regular general-purpose computers.
I currently use NREL’s 8-petaFLOPS Eagle supercomputer, considered the most energy-efficient data center in the world, to simulate and optimize traffic congestion problems for various US Department of Energy projects. I also run regional-scale traffic simulations using parallel computing on the Cori supercomputer, which has about 30 petaFLOPS performance and was the most powerful supercomputer in the world in 2017.
What are some of the challenges you have faced in taking this path?
The challenge with HPC is that there is so much to learn and many topics of interest, including parallel computing, programming and running jobs on HPC systems, and debugging and profiling HPC software. The HPC field is constantly evolving as new computing technologies and architectures emerge.
Most recently, the use of graphics processing units (GPUs) in HPC systems have enabled accelerating computing to higher performance than before, but requires knowing how to program and use GPUs in HPC systems. Attending the annual supercomputing conferences is one of the ways I keep up with new advances in HPC and get training in advanced HPC topics.
Any mentors you would like to thank?
I would like to thank my postdoc advisor, Dr. Sherry Li, a senior scientist at LBNL, who gave me my first training in HPC. Using her extensive HPC experience, she taught me the basics of HPC and showed me how to run my first job on Cori.
She also guided me to resources and classes to advance my HPC skills. Though I did not have a deep background in HPC, she believed that I could learn and successfully conduct my postdoc research.
By sharing her career path with me, my mentor Dr. Ann Almgren enabled me to believe that I could be successful as a research scientist.
I also want to thank Dr. Ann Almgren, another senior scientist at LBNL, who was my mentor as postdoc. By sharing her career path with me, she enabled me to believe that I could be successful as a research scientist. She answered questions I had about a research career and coached me as I was preparing to interview for my current research position at NREL.