Computer scientist Elsa Gonsiorowski joined Livermore Computing (LC) in 2016 and has steadily improved the Laboratory’s high-performance computing (HPC) ecosystem ever since. Her efforts affect applications, software, and users alike. With a Ph.D. in computer science from Rensselaer Polytechnic Institute, Elsa’s expertise includes I/O (input/output) operations—the lightning-fast communications between supercomputers and data storage devices—and scalable checkpointing as scientific applications continuously write data to storage.

For example, when the Laboratory deployed the Sierra supercomputer in 2018, Elsa contributed to acceptance testing and trained application teams on the system’s burst buffers, which provide storage layers between compute nodes and long-term data storage. In 2019, she was a member of the R&D 100 Award–winning team behind the SCR software, which reduces overhead on HPC systems by performing multilevel checkpoints to the parallel file system.

Elsa’s systems-level experience now informs her work supporting users through LC’s hotline, where she leads the technical consulting team. These consultants know the ins and outs of a wide range of HPC machines including the exascale El Capitan system. “LC is responsible for so many systems,” Elsa points out. “We have to find the common issues among 30-plus user-facing systems and a vast number of tools and infrastructure. In LC, we’re very knowledgeable about these systems and services, and we bring that expertise to meet users’ needs.”

Besides staying on top of this technical knowledge, Elsa spearheads much of LC’s user documentation, making sure instructions are up to date as technologies and policies change. Her team must also adapt to users’ preferences for contacting the hotline via phone, email, chat, or service ticket. “With queries coming in from all directions, any day can be busy trying to respond to every new issue that arises,” she says. “The gap between what the HPC vendors deliver and what the users are trying to do is much larger than you might think.”

Any hotline query could be indicative of a larger problem. Elsa explains, “Every issue is unique, so it’s hard to tell if you’re only looking at trees and not seeing the forest.” This challenge is part of the draw because, she notes, “It’s exciting that we have a breadth of people doing important work on these systems. I get to help them do it.”

Often consultants are the main communication conduit between system administrators and users, relaying alerts about unexpected system downtime or information about a new data service. This direct contact with users has strengthened Elsa’s dedication to software sustainability and developer productivity, both of which help users downstream.

A decade in, Elsa has learned the value of mentoring and community engagement. She has organized several conference workshops and currently serves as coordinator for the Better Scientific Software Fellowship, a program jointly sponsored by the Department of Energy and National Science Foundation. Once part of LC’s new employee cohort—affectionately named NEWTs for New Employee Welcome and Training—she is now among the NEWT Elders who continue to offer guidance and share best practices with the next generation. She states, “I want to help LC as a team uphold our values and fulfill our mission of delivering these amazing HPC systems, making them useful to users, and helping users do the best science they can.”

—Holly Auten