At the 2024 Supercomputing Conference (SC24), Hewlett Packard Enterprise (HPE) announced a historic breakthrough in computing.
The company delivered El Capitan, the world’s most powerful supercomputer, to the Lawrence Livermore National Laboratory (LLNL).
El Capitan operates at an extraordinary 1.742 exaflops. That means it can perform over 1.7 quintillion calculations per second. Its groundbreaking design makes it among the top 20 energy-efficient systems on the Green500 list.
This innovation is no accident. It represents years of collaboration between HPE, AMD, and the U.S. Department of Energy. Together, they designed a system that combines raw power with unmatched energy efficiency.
El Capitan’s cooling system sets it apart from other supercomputers. It uses a 100% fanless direct liquid-cooling system. This innovation reduces energy consumption and helps it achieve 58.89 gigaflops per watt.
HPE’s Cray EX platform powers El Capitan’s immense capabilities. AMD Instinct™ MI300A APUs provide the processing muscle. These chips integrate CPU and GPU cores with high-bandwidth memory in a single package.
The system features HPE Slingshot, a high-speed interconnect for seamless communication. Its more than 11,000 nodes can perform large-scale calculations simultaneously.
A custom near-node storage solution enhances data accessibility and reduces latency.
El Capitan’s primary mission focuses on U.S. national security. The National Nuclear Security Administration will use it to manage the nation’s nuclear stockpile. It will also support modernization efforts, ensuring safety and reliability.
The system enables highly detailed 3D modeling and simulations. These tools are crucial for addressing nuclear challenges. They ensure the U.S. maintains its technological edge in a competitive global environment.
LLNL’s Weapon Simulation and Computing Program Director, Rob Neely, praised the system’s potential. He explained how El Capitan combines modeling, simulation, and AI capabilities in a single platform.
This flexibility makes it a critical tool for advancing the lab’s national security mission.
The supercomputer will also address other security concerns, supporting nuclear nonproliferation and counterterrorism efforts. These capabilities add to its already impressive list of applications.
El Capitan is not limited to classified projects. It will drive research in energy security and climate change. The system will also play a role in power grid modernization and drug discovery.
Researchers can use El Capitan’s AI capabilities to solve complex problems. It will enable large-scale AI training and inference tasks. This combination of simulation and AI makes it a versatile and forward-looking platform.
Climate modeling will benefit from El Capitan’s computational power. Researchers can refine models and create more accurate predictions. These advancements will help mitigate the effects of climate change.
In the medical field, El Capitan will accelerate drug discovery. Its ability to simulate molecular interactions can speed up the development of new treatments, and the insights gained could lead to life-saving breakthroughs.
Energy research will also see dramatic improvements. El Capitan can analyze power grid data at unprecedented speeds. This capability will support the transition to renewable energy and enhance grid reliability.
El Capitan’s energy efficiency stands out as a critical achievement. Its design prioritizes sustainability without sacrificing performance. The liquid-cooling system exemplifies this approach by minimizing heat and reducing power consumption.
HPE, AMD, and LLNL spent five years developing this system. Their collaboration involved cutting-edge technology and innovative solutions. The result is a supercomputer that balances power, versatility, and efficiency.
The system relies on a Lustre-based global file system. This setup allows tiered, dynamic data access across all compute nodes, ensuring researchers can efficiently handle massive datasets.
AI integration makes El Capitan a unique resource. It bridges the gap between traditional simulations and machine learning. This dual focus enables faster, more accurate results in both domains.
The supercomputer also supports fusion research at the National Ignition Facility, providing the computational power needed for high-energy-density science.
These studies could lead to breakthroughs in energy production.
El Capitan’s impact extends far beyond its technical specifications. It represents a commitment to innovation and progress. By addressing critical challenges, it sets the stage for transformative discoveries.
HPE’s leadership in exascale computing continues to redefine possibilities. El Capitan joins Frontier and Aurora as part of HPE’s portfolio. These systems collectively demonstrate the company’s dominance in the field.
The deployment of El Capitan underscores the importance of public-private partnerships. HPE, AMD, and the DOE worked closely to achieve this feat. Their collaboration serves as a model for future technological advancements.
Trish Damkroger, HPE’s Senior Vice President for HPC & AI Infrastructure Solutions, called the project a success. She emphasized the system’s potential to strengthen national security and enable scientific breakthroughs.
Her optimism reflects the broader impact El Capitan will have on multiple industries.
The future of exascale computing looks promising. Systems like El Capitan push the boundaries of what’s possible. They enable researchers to tackle the world’s most pressing challenges with unprecedented tools.
El Capitan is more than just a machine. It is a symbol of innovation, collaboration, and progress. As it begins operation, the world watches, anticipating the discoveries it will unlock.