Lawrence Livermore National Laboratory’s El Capitan supercomputer is poised to become the world’s fastest system. Learn about the process of bringing this world-leading computing marvel to life.
For over 70 years, Lawrence Livermore National Laboratory (LLNL) has applied science and technology to enhance U.S. security and global stability. LLNL research covers vital security-related areas, including nuclear deterrence, defense, and energy. The technology foundation powering this critical research is undergoing significant enhancements as LLNL is partnering with HPE and AMD to stand up one of the world’s most powerful supercomputers, El Capitan.
This cutting-edge system is expected to offer significantly higher capabilities to LLNL, delivering at least 16 times faster performance than their current flagship supercomputer. It will address essential issues ranging from nuclear stockpile stewardship to threats posed to the environment to healthcare research.
In the words of Judy Hill, computational science project leader at LLNL, “We expect El Capitan to be an engine for artificial intelligence and deep learning… it will enable scientific breakthroughs that we can’t even imagine.”
Check out this brief video for a summary of the system:
Path to El Capitan
In early 2020, LLNL announced the selection of HPE and AMD as technology partners to build El Capitan. Since then, these organizations have collaborated closely, laying extensive groundwork necessary in anticipation of an intended go-live date several years in the making. From a full facility modernization project to hardware deliveries, software configuration, comprehensive testing, and more, this has been a highly complex undertaking. You can get all the details in a series of blog posts published by the LLNL team. Here’s the series: Road to El Capitan.
The resulting computing power of El Capitan is expected to exceed two exaflops, which translates to an astounding two times a quintillion calculations per second. To understand the magnitude of this power, consider what this means: If all 7.7 billion people on Earth each completed one calculation per second, it would take over eight years to do what El Capitan can do in 1 second. Let’s look at the building blocks that make this speed and power possible.
A solid technology foundation to accelerate discovery
El Capitan is built on HPE’s AI-native architecture, combining leading supercomputing technologies and services, including:
HPE Cray Supercomputing EX4000: a liquid-cooled blade-based, high-density supercomputer designed from the ground up to deliver the utmost in performance, scale, and density with up to 64 compute blades per chassis.
HPE Cray Supercomputing EX255a Accelerator Blade: ideal for AI and HPC workloads, this supercomputing accelerator blade features the AMD Instinct™ MI300A APU. The APU combines CPU and GPU processors and high-bandwidth memory into a single package. By reducing the physical distance between different types of processors and creating unified memory, it enables fast data transfer speeds, impressive HPC performance, easy programmability, and great energy efficiency. This should reduce the energy required to operate El Capitan compared to previous-generation supercomputers.
HPE Slingshot interconnect, an open, Ethernet-based high-performance network that provides state-of-the-art small packet performance on standard links
AMD Instinct™ MI300A Accelerator APU nodes: APUs, or Accelerated Processing Units, combine CPU cores and GPUs, allow for efficient resource sharing, and can handle computing and graphic-intensive tasks simultaneously.
HPE Cray Supercomputing Storage in two layers: All flash local NVMe storage inside the compute chassis that is used for creating temporary Lustre filesystem instances with a lifetime of a single job and the persistent external global Lustre file system that is built from Cray ClusterStor E1000 Storage Systems.
HPE Services: From design, planning, development, installation, and operation, the HPE Services team works closely with AMD to help LLNL build El Capitan and implement their desired outcomes.
A proven partnership to power your AI environment
While your organization might not need exascale computing, you can benefit from exascale-class performance on a solid technology foundation to run the AI workloads of today and tomorrow. HPE’s AI-native architecture supports running AI models that typically run one workload across thousands of nodes, is designed to run both on-premises and in the cloud and is based around an open ecosystem that is sustainable by design.
HPE and AMD are accelerating innovation and discovery in the AI era with breakthrough supercomputing technology, proven to scale to the largest of requirements. Our solutions are engineered for extreme performance, flexibility, and efficiency to meet any requirement, accelerate data-intensive workflows, and deliver better outcomes faster.
El Capitan is the next of several notable supercomputers powered by HPE and AMD; others include:
Frontier, developed for the U.S. Department of Energy’s Oak Ridge National Laboratory (ORNL), is the world’s first exascale system and is currently #1 in the latest edition of the TOP500 list. It was recognized by Time magazine as one of the top inventions of 2023.
LUMI, located at CSC’s data center in Finland, is Europe’s fastest system and holds the #5 spot in the TOP500 and #7 in the GREEN500.
Hunter and Herder are being developed for the University of Stuttgart. When complete, Herder, an exascale system, will significantly expand Germany’s high performance computing capabilities. Read the announcement here.
As you progress toward an AI-powered organization, you can be confident in HPE’s ability to help you accelerate the journey with our proven and trusted supercomputing solutions, technology, and services.
Learn more
Meet Diana Cortes, Marketing Manager, Data Solutions
Diana has spent the past 25 years working with the technologies that power the world’s most demanding IT environments and is interested in how solutions based on those technologies impact the business. A native from Colombia, Diana holds an MBA from Georgetown University and has held a variety of regional and global roles with HPE in the US, the UK and Sweden. She is based in Stockholm, Sweden.
Compute ExpertsHewlett Packard Enterprise
twitter.com/HPE_Craylinkedin.com/showcase/hpe-servers-and-systems/hpe.com/supercomputing