After experiencing the loud bang and flash of a transformer on a pole in her backyard blowing up when she was seven, Anna Maria Bailey had always imagined using her electrical engineering degree to work for a utility power company building power plants or dams. Years later in 1987, another explosion got Bailey, working at the time for the Energy Division of the California Department of Water Resources, interested in the possibility of working for LLNL. There was a national media storm covering the explosion, which resulted in damage to a parking lot and nearby buildings and resulting in extensive power outages for both Lawrence Livermore and Sandia National Laboratories.
“It was all over the news, and I had friends who worked at the Labs at the time who said they went days without power. It was a big deal,” says Bailey, who is LLNL’s Chief Engineer for High Performance Computing (HPC) and a Distinguished Member of Technical Staff. “I wanted to know what had happened and how they would work to fix it.”
The resulting study on the Lab’s antiquated electrical system led to a wave of job postings for power engineers, so Bailey applied. Starting at the Lab in 1989, she was immediately involved in improving the site-wide electrical system starting with the main substation. Bailey was named design and construction manager for the project soon after in 1990 until its completion in 1998.
Bailey recalls getting her first glimpse into the world of HPC and the work that would become part of her legacy at LLNL in the form of a phone call from Mark Seager, who would go on to become chief technology officer for HPC, heading up the project they’d work on together.
“I hung up on him. I didn’t know who he was, but he had apparently seen an article in Newsline [the Lab’s internal news service] about my work at the Lab,” recalls Bailey. “I barely caught his name and where he was from before he began demanding I find a way to produce between 1.5 and 3 megawatts for some supercomputer he was purchasing—all in the span of a few months!”
At the time, Bailey could not imagine why a supercomputer would require so much power or how it would be done, and she still had progress to make on her primary project. But no sooner had she hung up on Seager than he was standing in her office doorway completely aghast at her abrupt ending to the conversation. He convinced her to speak with Michael McCoy, LLNL’s Accelerated Strategic Computer Initiative Program Manager.
“Michael went to the white board while Mark explained what they were embarking on and his vision for supercomputing at the Lab,” says Bailey. “He explained what a flop [floating-point operation] was and what HPC could do for the Lab and, more importantly, stockpile stewardship.” The sketch would eventually become Building 453, which is now home of the Sierra supercomputer and the upcoming El Capitan exascale system.
Bailey agreed to join the team, kicking off lifelong friendships and a project that would succeed at everything Seager and McCoy had aimed for in those early days and more.
In 2022, Bailey and her team finished the Exascale Computing Facility Modernization (ECFM) project, which began in 2017 and expanded Building 453 to 85 megawatts, far surpassing the original 3 megawatts Seager mentioned nearly 30 years before. Under Bailey’s direction, the project was completed in record time, 9 months ahead of schedule, and $9 million under budget. “Michael knew what was coming even though I couldn’t appreciate it at the time,” she notes. “It’s amazing to now stand on the shoulders of the work I was doing when I first came on.”
Bailey says the biggest challenge throughout these two main projects and across her career at the Lab has been the speed at which technology evolves. Starting in 1998, the Lab was depending on the White supercomputing system using commercial-grade facility equipment, but components of computers like Sequoia and Sierra have gotten much more complex, requiring industrial-level facility solutions. El Capitan will require even more power, operating on utility-level infrastructure when it comes online in 2024.
In her time at LLNL, Bailey is most proud of her projects’ excellent safety record and sees all of her work supporting the mission as her way of giving back to her country. She has enjoyed what she refers to as “many careers” at the Lab and the ability to jump around to follow her interests. She credits the success of her projects with the ability to work easily with a range of professionals towards the same goals.
“Whether you’re working on the systems, facilities, or the hardware side of Livermore Computing, everyone is on the same team,” Bailey states. “We really work more like a production plant than anything else which is very atypical for a lot of other organizations.”
Bailey is active in the international HPC community as co-lead of EEHPCWG, the Energy Efficiency High Performance Computing Working Group, a group of professionals worldwide collaborating to push the envelope of HPC with a focus on energy efficiency. For the last eight years, she has contributed to recruitment and retention for the Lab by forging a relationship with young engineers in her alma mater’s student-run engineering clubs.
When she’s not being a human powerhouse for Livermore Computing, Bailey likes to relax with her husband, who retired from the Lab in 2007 and now runs their winery. Between mentoring young winemakers, accepting awards for their 2017 cabernet sauvignon, and helping run the business, Bailey enjoys caring for her 91-year-old mother and making plans for “someday” travel from their home in the foothills of the Sierras. “It’s kind of ironic because HPC is all about being quick, how fast you can speed up processes,” said Bailey. “In winemaking, it’s the opposite. And sometimes it’s good to slow down.”