AMD And HPE Unveil El Capitan, World’s Most Powerful Supercomputer With Over 11M Cores

At the Supercomputing 2024 conference currently underway in Atlanta, AMD, in collaboration with HPE and the Lawrence Livermore National Laboratory (LLNL), announced what has been officially declared as the world’s most powerful supercomputer, El Capitan. El Capitan is the first exascale-class machine for the National Nuclear Security Administration (NNSA), but will also be powerful resource for the NNSA Tri-Labs- LLNL, Los Alamos, and Sandia National Laboratories as well.

El Capitan’s main area of focus is national security. The system will be tasked with the complex modeling and simulation required for the NNSA’s Stockpile Stewardship Program that certifies the aging nuclear weapons stockpile. It will, however, also handle the computing requirements for other critical nuclear security missions and counterterrorism research.

El Capitan was built by HPE on its direct liquid-cooled blade-based, high-density clustered Cray EX255a HPC system. It is powered by AMD Instinct MI300A APUs – which pack both CPU and GPU cores -- and leverages HPE’s high-bandwidth Slingshot-11 interconnect technology.

“We are thrilled to see El Capitan become the second AMD powered supercomputer to break the exaflop barrier and become the fastest supercomputer in the world. Showcasing the incredible performance and efficiency of the AMD Instinct MI300 APUs, this groundbreaking machine is a testament to the dedicated work between AMD, LLNL and HPE,” said Forrest Norrod, executive vice president and general manager, AMD. “At AMD, we are driving the future of computing with leadership performance and capabilities that will continue to define the convergence of HPC and AI for years to come.”

The AMD MI300A packs 24 ‘Zen 4’ cores across three chiplets, in conjunction with six accelerated compute dies (XCDs) with 38 compute units (CUs), each with 32 KB of L1 cache, 4 MB L2 cache shared across CUs, and 256 MB AMD Infinity Cache shared between XCDs and CPUs. There’s also 128GB of HBM3 memory shared coherently between CPUs and GPUs, offering up to 5.3TB/s of peak throughput. All told, El Capitan features a whopping 11,039,616 total cores, offering a measured 1,742 PFlop/s (Rmax), with a theoretical peak 2,746.38 PFlop/s. And the system consumes 29,580.98 kW of power.

That’s a ton of computing power, which landed El Capitan at #1 atop the Top500 supercomputer list, making it the fastest supercomputer in the world at this time. The system also lands at number 18 on the on the Green500 list, which ranks the systems in the Top500 in terms of energy efficiency. AMD has had great success in its supercomputing efforts. The company powers the no. 1 and no. 2 fastest supercomputers in the world, in El Capitan and Frontier (Oakridge National Labs), but also 5 of the top 10 on the Top500 and 40% of the ten most energy efficient supercomputers in the world.