
Exascale supercomputer marks ‘new era’ in AI, science research
The Frontier supercomputer, says the company, has reached 1.1 exaflops, making it the world’s first supercomputer to break the exascale speed barrier, and the world’s fastest supercomputer, according to the Top500 list of world’s most powerful supercomputers. Frontier also ranked number one in the mixed-precision computing category, which rates performance in formats commonly used for artificial intelligence, with a performance of 6.88 exaflops.
Additionally, says the company, the new supercomputer claimed the number one spot on the Green500 list as the world’s most energy efficient supercomputer with 52.23 gigaflops performance per watt, making it 32% more energy efficient compared to the previous number one system.
“Today’s debut of the Frontier exascale supercomputer delivers a breakthrough of speed and performance, and will give us the opportunity to answer questions we never knew to ask,” says Justin Hotard, executive vice president and general manager, HPC & AI, at HPE. “Frontier is a first-of-its-kind system that was envisioned by technologists, scientists and researchers to unleash a new level of capability to deliver open science, AI and other breakthroughs, that will benefit humanity.”
Built with HPE Cray EX supercomputers, Frontier, says the company, will speed up discoveries, make breakthroughs, and address the world’s toughest challenges, and is expected to have significant impact in critical areas such as cancer and disease diagnosis and prognosis, drug discovery, renewable energy, and new materials to create safer and sustainable products. The supercomputer, which is more powerful than the next top seven of the world’s largest supercomputers combined, will allow scientists to model and simulate at an exascale level to solve problems that are 8X more complex, up to 10X faster. Frontier is also expected to reach even higher levels of speed with a theoretical peak performance of 2 exaflops.
In addition to modeling and simulating complex scientific research, across biological, physical and chemical sciences, with higher resolution, Frontier will also enable dramatic breakthroughs in AI. At an exascale speed, Frontier’s users can develop AI models that are 4.5X faster and 8X larger, allowing to train more data that can increase predictability and speed time-to-discovery.
The company designed and built Frontier with the following state-of-the-art technologies, delivered through the HPE Cray EX supercomputers, to offer dramatically higher performance to model and simulate at a new level, and target new applications in AI and machine learning to increase accuracy faster and more efficiently:
- 74 HPE Cray EX cabinets each weighing more than 8,000 lbs. Each node contains one optimized 3rd Gen AMD EPYC processor and four AMD Instinct MI250x accelerators, for a total of 9,408 CPUs and 37,632 GPUs in the entire system
- 90 miles worth of HPE Slingshot networking cables, which delivers the world’s only high-performance Ethernet fabric designed for next-generation HPC and AI solutions. These include larger, data-intensive workloads, to address demands for higher speed and congestion control for applications to run smoothly and boost performance.
- Cray Clusterstor E1000 storage system, which enables Frontier’s Orion storage system to deliver 700 petabytes of storage capacity, peak write speeds of more than 35 terabytes per second, and more than 15 billion random-read input/output operations per second.
- Sophisticated liquid-cooling capabilities that also promotes a quieter datacenter, compared to a noisier, air-cooled system, to efficiently remove heat from high power devices such as processors, GPUs and switches, through an auxiliary Cooling Distribution Unit (CDU).
In addition to Frontier, the company is also delivering exascale supercomputers to the U.S. Department of Energy’s Argonne National Laboratory and Lawrence Livermore National Laboratory.
