Edgar Leon named IEEE senior member

Edgar Leon, a computer scientist in the Livermore Computing (LC) Division, has been elevated to the grade of Institute of Electrical and Electronics Engineers (IEEE) senior member.

Only 7 percent of IEEE's approximately 431,000 members hold this prestigious status, which reflects professional maturity and requires extensive experience and documented achievements. 

"I’m honored to be named a senior member and I am most grateful to my LLNL, Sandia and colleagues at the digital architecture company ARM who supported my application," said Leon, who, in addition to his pioneering research and development in high performance computing (HPC), relishes serving as a mentor to computer science students and postdocs.

Working on world-class supercomputers at a U.S. national lab is not what Leon, a native of Mexico, envisioned when he began his scientific studies in preparation for university. He originally set out on a path to medical school strongly encouraged by his family. But as part of his high school preparatory studies, he pursued computing and earned an associate degree in computer science that would turn out to be fortuitous.

Leon got to embrace his love of computing after not gaining admission to Mexico’s top medical school. "I was very interested in what you could do with programming and how that could be applied to scientific problems," Leon said. "A part of my personality is great curiosity about why we do things a certain way.

"I feel really lucky that I ended up in computing," he said.

Leon began his undergraduate studies at UNAM University in Mexico and completed his senior year at the University of New Mexico in Albuquerque where he continued his studies, earning his master’s and doctoral degrees in computer science.

He was a member of the Novel Systems Architecture Department at IBM Research, where he contributed to the development of the IBM Power7-IH system, which ranked 10th in the top 500 supercomputers in the world (November 2012). 

Leon has spent his career confronting the daunting challenges in the development of ever more powerful high performance computing systems. He helped prepare for Sequoia, the 17.5 petaflops (quadrillions of operation per second) IBM Blue Gene/Q supercomputer ranked No. 3 in the world on the Top500 list, and now is laying the groundwork for Sierra, the 150 petaflops IBM system scheduled for delivery in 2017.

The principal challenge is ensuring the "resilience and reliability" of complex HPC systems with tens of millions of components that operate under great power constraints, Leon says. "The more components you have, the more failures occur.

"The question in deploying exascale supercomputers is how our applications can take advantage of these new capabilities," he said.

Leon's pioneering contributions include a mechanism to simulate novel computer architectures in large systems and the application of "cache injection" to high performance computing. The latter contributed to the commercialization of cache injection, a technique for tolerating memory latency, or slow data access, in HPC systems.   

His research at LLNL is part of a larger DOE/NNSA effort to develop and leverage exascale computing capabilities to ensure the U.S.’ global leadership in high performance computing. In particular, Leon works on performance and reliability of communication libraries at scale, emerging memory technologies for exascale systems, power-aware computing strategies and resilience of future systems.

His investigation in resilience of extreme-scale systems includes innovations to cope with runtime errors using data reconstruction techniques at the application level. Leon's work in power-aware computing provides application developers insights into the best program optimizations when tuning code for future supercomputing platforms in terms of power, energy and performance. In addition, Leon is researching new methods to address system noise, which affects the performance of numerous scientific applications. He also is developing mechanisms to utilize multi-level memory systems more effectively on emerging architectures.

Leon said he finds his LLNL assignment profoundly satisfying because it provides a balance of applied scientific research and work with people. "Here at Lawrence Livermore, we’re driven by missions and not by commercial products, yet we work with top people from academia and industry as well as other labs," he said.

Remembering his days as a student and the mentors who helped him in his career, Leon dedicates time and energy to working with students. "The challenge for a mentor is to identify and leverage the unique skills each student brings to an internship. Not everyone is the same," Leon says. "Students are excited by the magnitude of the problems they work with here and, at the same time, we can have an impact on the focus of their studies and their dissertations."

"We’re trying to create a pipeline of computer scientists for the national labs," he says. "It’s not that long ago that I was going through this process."

For Leon, the pipeline to LLNL was Tony Baylis, the Lab’s director of diversity programs. They met at a conference where Baylis was recruiting computing talent. The people aspect of LLNL culture emphasized by Baylis was one of the attractions that brought Leon to Livermore.

"It’s important to me to have an impact on people, not just science," Leon says.