Lessons from the Frontlines of HPC: Dona Crawford on Leadership and Discovery

Dona Crawford

Former Associate Director for Computation, Lawrence Livermore National Laboratory

Fostering Innovation at Scale

SC25 Communications Team: Hi, Dona. As someone who has led thousand-person teams driving breakthrough computing capabilities, what leadership principles have you found most crucial for fostering innovation in large-scale technical organizations?

Dona Crawford: I believe that leadership is everyone’s business. It’s not about job titles or hierarchy; it’s about how you show up, how you engage, and how you contribute to something bigger than yourself. No matter where you are in an organization, you have a role to play.

For me, leadership starts with clarity—clarity of purpose, clarity of role, and clarity of who you’re serving. At Lawrence Livermore National Laboratory, our purpose in Computation was to build and sustain an integrated computing environment that could support everything from blue-sky research to mission-critical applications. That meant hardware, operating systems, scalable libraries, visualization tools, data analytics, file systems—everything needed to turn raw computing power into scientific breakthroughs. It wasn’t just about having the fastest machines; it was about advancing the entire ecosystem.

The most effective leaders aren’t the ones with all the answers—
they’re the ones who know how to bring out the best in others.

DONA CRAWFORD

My role wasn’t to be the technical expert. It was to find the right people, create an environment where they could thrive, and remove barriers so they could do their best work. The most effective leaders aren’t the ones with all the answers—they’re the ones who know how to bring out the best in others. That’s why I believe in being open to the outcome. If you’re too locked into a single solution, you risk missing a better one. Leadership is about direction, yes, but it’s also about listening, adapting, and making space for innovation to happen.

One of the most important things I’ve learned is the power of focus. When I’m working on something, I give it my full attention. I like to call it showing up and being present. It’s easy to get pulled in a hundred different directions, especially in a fast-moving, high-stakes environment like LLNL. But you can’t do everything at once. The key is to be fully engaged in whatever you’re doing at that moment, and then shift your attention completely when it’s time to move on to the next thing. That kind of presence isn’t just good for productivity—it’s good for leadership. Because when you’re fully present, you make better decisions, build stronger relationships, and create a culture where others feel seen and heard.

At the end of the day, leadership is about service. It’s about understanding the mission, supporting the people who make it possible, and staying open to the best path forward—even when it’s not the one you originally had in mind.

AI and HPC: The New Frontier

SC25: AI and HPC are rapidly evolving together. What breakthroughs do you see emerging from this convergence, and how will they transform scientific research?

Dona Crawford: The convergence of AI and HPC is revolutionizing science. We’re witnessing an unprecedented acceleration in data analysis, simulations, and discovery—unlocking insights that were previously out of reach. The question isn’t just about what AI and HPC will impact, but what won’t they?

One of the most exciting things is how AI and HPC are transforming drug discovery. By screening vast molecular libraries computationally, researchers can accelerate drug development dramatically. The ability to analyze large-scale genetic datasets is opening the door to personalized medicine, where treatments are tailored to individual patients.

And perhaps most excitingly, 3D bioprinting of patient-specific bones and organs is no longer science fiction. This one is personal for me—after I retired from LLNL, I worked with the Organ Donor Association of Northern California to explore how bioprinting could one day eliminate the need for organ donors altogether. Imagine a world where no one has to wait for an organ transplant—where an organ could be developed just for them. This technology is already in early stages with human tissue, and as it evolves, it could completely redefine treatment strategies for injuries and diseases. While we’re not printing fully functional organs yet, the field is advancing rapidly. That’s the kind of future AI and HPC are making possible. 

Aside from biology, we are entering a new era for materials science. Of course, materials have competing properties—strength, flexibility, durability, weight, heat resistance, conductivity, and manufacturability. Traditionally, optimizing these trade-offs was painstakingly slow. But now, AI and HPC allow us to computationally design entirely new materials by pinpointing ideal properties on an Ashby diagram and simulating atomic-level interactions. We’re no longer limited to what naturally exists—we can engineer materials that have never been seen before.

Dona Crawford stands in front of the Ford-Magna Lightweight Concept vehicle, a project that leveraged LLNL’s high performance computing and materials modeling to push the boundaries of automotive lightweighting and fuel efficiency.

There are so many areas worth mentioning, but I’ll give you one more. Climate science is an area that is very important. AI-driven HPC is reshaping climate modeling and weather forecasting by enabling the analysis of massive datasets at finer resolutions than ever before. We can now simulate energy and matter flows through the climate system on smaller 3D grids, vastly improving predictions of extreme weather, long-term climate trends, and mitigation strategies. With climate change driving more frequent and intense weather events, these advancements are becoming critically important.

And these are just a few examples. Honestly, there’s no field that won’t be transformed by the combination of AI and HPC. With the help of these tools, we are unlocking solutions to problems that were once thought unsolvable.

HPC’s Role in Transforming Lives

SC25: Building on these technical advances, through your work at LLNL and now the Livermore Lab Foundation, you’ve seen HPC tackle increasingly complex societal challenges. Which emerging applications do you see as having the most transformative potential for improving people’s lives?

Dona Crawford: When we think about HPC, we often picture simulations for national security or fundamental physics. But some of the most transformative applications are unfolding in ways that will deeply affect our daily lives—especially as we age.

One area I’m particularly excited about is brain mapping for neurodegenerative diseases. As lifespans extend, so does the prevalence of conditions like Alzheimer’s, Parkinson’s, ALS, and multiple sclerosis—diseases that progressively impair brain function. If we can create high-resolution maps of the human brain, identifying individual neurons and their intricate connections, we can gain unprecedented insights into what triggers these diseases, how they progress, and ultimately, how to treat or even prevent them.

This is a monumental challenge. It requires analyzing massive datasets from multiple imaging modalities—fMRI, EEG, diffusion tensor imaging—while running complex simulations of neurons and reconstructing three-dimensional models of brain function. AI and HPC together make this possible, giving us the computational power to integrate vast amounts of data and detect patterns beyond human perception. And, frankly, I hope they hurry up—because we’re all getting older by the day!

Another area I often think about is autonomous vehicles. For many older adults, the ability to drive represents independence. But as reflexes slow and vision declines, driving safely becomes more difficult. The promise of fully autonomous vehicles could revolutionize mobility for aging populations, allowing people to maintain their freedom without compromising safety. However, for this technology to become truly viable, we need real-time, ultra-reliable navigation systems that can react instantly to the unpredictable nature of driving. AI and HPC are at the heart of making that a reality.

These are just two examples, but the broader convergence of AI and HPC is reshaping how science itself is done. The ability to process, simulate, and analyze massive datasets at unprecedented speeds is accelerating discoveries in nearly every field—from medicine and climate science to national security and economic development. As Eric Schmidt, Google’s former CEO, put it: “Science is about to become much more exciting—and that will affect us all.”

I couldn’t agree more.

The Keys to Success in HPC Projects

SC25: When you consider these technological transitions and other ambitious HPC projects you’ve led, what factors have you found most critical in determining a successful outcome?

Dona Crawford: People, people, people. That’s the number one factor. You can have all the computing power in the world, but without the right people—motivated, dedicated, creative, and smart—you won’t get anywhere.

You can have all the computing power in the world, but without the right people—motivated, dedicated, creative, and smart—you won’t get anywhere

DONA CRAWFORD

Beyond that, success comes down to alignment and execution. You need a clearly defined scope and goals that everyone—your team, your customers, your funders, your partners—buys into. If people aren’t aligned from the start, the project is in trouble before it even begins.

Resources matter too, and not just money. You need access to the right tools—computing time, data, and prior work to build on. Reinventing the wheel slows everything down.

And then there’s timing. If a project moves too fast, quality suffers. If it drags on, costs go up, people lose focus, and suddenly you’re reworking things that should have been settled. It’s about finding the right balance—long enough to do it right, but not so long that it loses direction.

Clear decision-making keeps things on track. Everyone needs to know who’s responsible for what, and leadership has to be decisive. If no one’s empowered to make the tough calls, you’ll waste time spinning in circles.

Dona Crawford gives former California State Assembly Member Catharine Baker a tour of Lawrence Livermore National Laboratory’s high performance computing facilities, discussing the impact of supercomputing on scientific research and national security.

Finally, check in often. Progress reviews aren’t just about reporting—they’re about catching problems early before they spiral. The sooner you tackle an issue, the better your chances of keeping things on course.

It’s pretty simple: get the right people, give them a clear mission, support them with the resources they need, and stay engaged. Do that, and your chances of success go way up.

SC: The Heartbeat of HPC

SC25: The SC conference has been a constant throughout your career, including your role as SC97 chair. Looking at today’s rapidly evolving computing landscape, what makes SC more essential than ever for our community?

Dona Crawford: SC has always been more than just a conference—it’s the heartbeat of the HPC community. I attended my first SC in 1988—the very first one—and even then, it was an unparalleled gathering of scientists, engineers, educators, developers, and industry leaders. Over the years, it has only grown in scale, influence, and importance.

What makes SC essential today is the sheer pace of technological change. HPC is evolving faster than ever, with AI, quantum computing, and exascale architectures reshaping what’s possible. SC is where the community comes together to stay ahead of emerging trends, connect with experts across academia, industry, and government, and collaborate on the ideas that will drive the next wave of innovation. It’s a place where the best minds push the boundaries of what computing can do.

The theme for SC25, HPC Ignites, is especially fitting. SC has never been just about keeping up with progress—it’s about shaping what comes next. The relationships and conversations that happen here fuel breakthroughs long after the conference ends. That’s why SC remains a must-attend event for anyone serious about the future of HPC.

I’m also thrilled that SC25 is being chaired by Lori Diachin, someone I deeply admire and respect. And here’s a little fun fact: besides being an extraordinary technical leader, she’s also an incredible artist!

A Lifetime of Memories at SC

SC25: Finally, thinking back on all your years at SC, is there a moment that stands out as a turning point for the industry or a great example of the community coming together?

Dona Crawford: When I first started in computing in the late 1970s, we were at the tail end of the mainframe era—a period that had lasted more than 30 years. Transitioning to vector computing in the 1980s was disruptive, but manageable, thanks to the fact that the Cray-1’s scalar processor was faster than that of its predecessor, the CDC 7600. That shift ushered in a 15-year period of stability.

So one of my most vivid SC memories is from the very first SC conference in 1988, when Seymour Cray himself was our keynote speaker. He talked about the Cray X-MP and Y-MP, and there was an unmistakable sense of momentum in the room—like we were standing at the edge of something big.

Dona Crawford and Professor Yutong Lu stand in front of Tianhe-2, one of the world’s most powerful supercomputers, highlighting the global collaboration in high performance computing that drives scientific discovery and innovation.

Then came the shift to massively parallel processing (MPP) and distributed memory systems in the mid-to-late 1990s, and that was a much harder transition. Suddenly, we had to rewrite code for cache-based commodity processors instead of vector processors and scale up to thousands of processors. It was a challenge, but the community responded—the development of the Message Passing Interface (MPI) and ASCI’s funding of 3D scientific codes helped HPC keep moving forward. That period lasted less than a decade before we entered what is still an ongoing transformation: multi-core and heterogeneous computing.

Since 2008, we’ve seen an explosion of architectures—specialized processors, chiplets, GPUs, FPGAs—each adding complexity. Meanwhile, cloud providers are pouring vast resources into computing infrastructure, but their priorities don’t always align with traditional HPC’s focus on physics-based, multi-scale simulations. The future of HPC remains uncertain in many ways because it lacks the same level of investment as the cloud industry.

Yet, through all these changes, SC remains the place where the HPC community gathers to understand what’s next. We come together not just to discuss where we’ve been but to define where we’re going. SC has always been the conference where the big shifts become clear—where we see the industry not as it is, but as it will be.

SC25: Do you have a favorite memory from attending SC over the years?

Dona Crawford: Oh, man, I have too many—really, too many!

SC has brought together so many incredible people over the years. The fact that I could talk to Bill Gates, Burton Smith, and Seymour Cray—where else would I have had that opportunity?

Back then, I was just a staff member, not yet leading computing at Livermore Lab. But at SC, I had direct access to the rock stars of our field. That, for me, has to be among my favorite memories—being able to have those conversations and learn from the best.

Join Us at SC25 in St. Louis

Collaboration and continuous learning are essential to advancing high performance computing. SC25 offers an unparalleled opportunity to expand your knowledge, connect with experts, and engage with the global HPC community.

Attendees will experience cutting-edge research and real-world applications through technical presentations, papers, workshops, tutorials, posters, and Birds of a Feather (BoF) sessions—all designed to showcase breakthroughs in AI, HPC, and emerging technologies. The conference provides a unique platform where leaders from industry, research institutions, and academia come together to share insights and innovations shaping the future.

Join us for a week of discovery and collaboration at SC25 in St. Louis, November 16–21, 2025.

Registration opens in July—don’t miss the chance to be part of the world’s premier supercomputing conference!