The fastest and most energy efficient supercomputers in the world are HPE-built systems
HPE maintains the top position in performance share of the world’s fastest supercomputers, delivering next-generation capabilities in high performance computing workloads to accelerate scientific discovery
- Two HPE-built exascale supercomputers verified as the world’s most powerful
- HPE has developed four of the top 10 fastest systems in the world and the majority of the top 10 most energy efficient supercomputers
- New leadership-class systems include the Venado system for Los Alamos National Laboratory and Isambard-AI for the University of Bristol
Photo credit: Argonne National Laboratory
HPE delivers two exascale supercomputers – the fastest in the world
The latest TOP500 list is out, and Hewlett Packard Enterprise (HPE) has many reasons to celebrate along with its customers. The 63rd edition of the twice-yearly list that verifies and ranks the world’s most powerful supercomputers revealed that four of the top 10 systems were built by HPE and two of them are exascale machines.
Topping this list and ranking No. 1 for the fifth time, Frontier, the HPE-built exascale system for Oak Ridge National Laboratory (ORNL) remains the fastest supercomputer in the world. Frontier was the first verified exascale supercomputer in the world and has won several accolades for its ability to accelerate scientific discovery by running larger models and simulations than previously possible, advancing drug discovery, cancer and disease diagnosis, renewable energy, and materials science.
The 63rd edition of the twice-yearly list that verifies and ranks the world’s most powerful supercomputers revealed that four of the top 10 systems were built by HPE and two of them are exascale machines
Aurora, an HPE Cray EX supercomputer for Argonne National Laboratory (ANL), has been verified as the second system to reach exascale. Ranking No. 2 on the TOP500, Aurora has achieved a performance of 1.012 exaflops. More information about Aurora which was built in partnership with Intel, the U.S. Department of Energy, and ANL can be found here.
The momentum doesn’t stop there. LUMI, built for the EuroHPC JU and hosted by the LUMI consortium and located at CSC in Finland, came in at No. 5. Researchers are utilizing LUMI to model climate change and early cancer detection by training an AI model on millions of samples to produce more accurate diagnoses.
The Alps supercomputer, built for the Swiss National Supercomputing Center (CSCS) debuts as a top 10 system, ranking No. 6. Based on the HPE Cray EX supercomputer NVIDIA GH200 Grace Hopper Superchips, Alps achieved a performance of 270 petaflops. Researchers from a wide range of disciplines will be able to use the outstanding performance of Alps and its unprecedented AI capability to their advantage. Climate and weather, astrophysics, computational fluid dynamics, life sciences, quantum chemistry and particle physics will all benefit from Alps’ innovative architecture.
In addition to the top 10, Lawrence Livermore National Laboratory (LLNL) submitted three new systems to the TOP500 list from its El Capitan contract, which feature HPE Cray EX supercomputers with AMD Instinct™ MI300A APUs: El Capitan early delivery system (No. 46), RZAdams (No. 47), and Tuolumne (No. 48). Researchers at LLNL are running fusion ignition simulations on RZAdams through a project named “ICECap,” where AI techniques are being used to improve large multi-physics simulations. Tuolumne will support unclassified projects in fields like AI-backed computational drug discovery, bioresilience, climate modeling, astrophysics and other open science applications.
New systems on the top 10 of the Green500
In addition to developing powerful, exascale supercomputers, HPE continues to help its customers build some of the world’s most energy efficient systems – accounting for seven of the top 10 on the Green500.
The University of Bristol’s Isambard-AI system is new to the Green500 list. HPE recently installed phase one of what is projected to be the UK’s most powerful AI supercomputer. Consisting of a single rack of HPE Cray Supercomputing EX2500 featuring 168 NVIDIA Grace Hopper Superchips, Isambard-AI phase one is already one of the world’s most energy-efficient supercomputers, ranking No. 2 on the top 10 of the Green500 list. This system will offer researchers revolutionary computing power to make AI-driven breakthroughs in robotics, big data, climate research, and drug discovery.
Helios, which HPE recently delivered to Academic Computer Centre Cyfronet of the AGH University of Krakow, was recognized as the No. 3 most energy efficient supercomputer on the 23rd installment of the Green500, which verifies systems around the globe. Based on HPE Cray technology, Helios is Poland’s fastest supercomputer. More information about Helios, which features, 4th Gen AMD EPYC™ processors, NVIDIA GH200 Grace Hopper Superchips and NVIDIA H100 Tensor Core GPUs can be found here.
HPE continues to deliver a majority of the world’s top 10 most energy-efficient supercomputers, which also includes Alps (No.5), Frontier Test & Development System (No. 7), Venado (No. 8), Adastra (No. 9), and Setonix (No. 10). HPE is the leader in energy-efficient supercomputing by minimizing power consumption through the use of direct liquid cooling technology.
HPE is the leader in energy-efficient supercomputing by minimizing power consumption through the use of direct liquid cooling technology
Delivering supercomputing advancements around the world
In April, HPE joined representatives from NVIDIA, the U.S. Department of Energy, National Nuclear Security Administration and Los Alamos National Laboratory (LANL) to unveil the lab’s newest supercomputer, Venado. The system, delivered to LANL in March, is the result of a co-design process that highlights the power of public-private partnerships in continuing to advance scientific breakthroughs. Venado is an HPE Cray EX supercomputer featuring NVIDIA GH200 Grace Hopper Superchips.
In addition, HPE will continue working with customers who have recently announced new systems, including Eni, one of the world’s largest energy companies, and its upcoming HPC6 to be installed in Eni’s Green Data Center in Italy. HPC6 will be a key asset in meeting the challenge of achieving net zero, providing a crucial technological lever for Eni to gain competitive advantages in the development of new energy sources. In Germany, the University of Stuttgart will build two new supercomputers called Hunter and Herder at the High-Performance Computing Center (HLRS). Hunter is slated to be a transitional system while Herder is scoped as an exascale system that will significantly expand Germany’s HPC capabilities.
From ISC 2024 and Beyond
Supercomputing innovation continues to break down the barriers of what’s possible in scientific and medical research, fusion energy, climate forecasting and more. At ISC 2024, HPE celebrates the partnerships between public and private entities and collaboration among industry leaders which lead to groundbreaking achievements that will discover solutions to some of the biggest challenges society faces today.
Humanity’s ability to solve the world’s most complex challenges will be shaped by supercomputing-powered research accelerated by AI capabilities. ISC24 is a moment to celebrate the amazing achievements of HPE’s customers and partners as they drive scientific breakthroughs and innovation forward.