El Capitan, the record-breaking supercomputer built by HPE and housed at the Lawrence Livermore National Laboratory in California, has officially claimed the title of the world’s fastest supercomputer, according to the latest TOP500 rankings. The prestigious list, tracking supercomputing milestones since 1993, underscores El Capitan’s groundbreaking achievement.
“El Capitan marks another significant milestone in exascale supercomputing, bringing monumental performance, energy efficiency, and the capabilities to accelerate AI-driven scientific discovery and make incredible breakthroughs to strengthen national security and unlock new opportunities in renewable energy,” Trish Damkroger, senior vice president and general manager, HPC & AI Infrastructure Solutions, said in a press release.
The TOP500 team clocked El Capitan at a Linpack score of 1.742 EFlop/s, a standardized measure of how quickly the system can solve a complex sequence of linear equations.
El Capitan supports the work of the U.S. National Nuclear Security Administration, including research on nuclear weapons and deterrents. The lab also uses the machine for scientific research on energy security, climate change, power grid modernization, and drug discovery.
The supercomputer uses AMD 4th-generation EPYC processors, the Cray Slingshot 11 network for data transfer, and a custom storage solution. The world’s number one supercomputer has 11,039,616 combined CPU and GPU cores. It runs at 58.89 gigaflops performance per watt.
“El Capitan’s introduction continues the capability advancement needed to sustain our stockpile without returning to explosive nuclear testing,” Jill Hruby, Department of Energy undersecretary for nuclear security and NNSA administrator, said in a press release.
In May 2024, El Capitan appeared on the TOP500 list at number 46.
SEE: Dell offered new infrastructure for enterprise AI applications at the international supercomputing conference held this week.
Behind El Capitan, here are the most powerful computers in the world, according to the TOP500 list:
Frontier increased its exaflops to 1.353 Eflop/s on the November list, but that was still insufficient to prevent El Capitan from knocking Frontier down from the top spot. Frontier uses 9,066,176 cores and Cray’s Slingshot 11 network. It is used to study materials, energy, nuclear power, and genetics, among other scientific research.
Aurora is built on HPE’s Cray EX — Intel Exascale Compute blade with Intel Xeon CPU Max Series Processors and Intel Data Center GPU Max Series accelerators. It maintained its benchmark score from earlier this year, at 1.012 Exaflop/s. The lab uses the supercomputer for scientific research, including climate, materials, fusion energy, and energy storage.
Eagle is a cloud-based supercomputer cluster offered through Microsoft Azure. Eagle uses 14,400 NVIDIA H100 GPUs and Intel Xeon Sapphire Rapids processors to reach a benchmark score of 561.2 PFlop/s.
HPC6 joined the top five this year, bringing 477.90 PFlop/s from its home with the Eni energy company in Italy. HPC6 can now claim the title of the fastest system in Europe. TOP500 notes HPC6 uses the same architecture as Frontier.
TOP500 shows AMD and Intel are the best choices for the world’s most powerful supercomputers. Five of the top 10 use AMD processors, while three use Intel. The remaining two use NVIDIA or a custom ARM-based processor. The Slingshot-11 interconnect seems to be the most popular choice for connectivity, used by seven of 10.
