Supercomputers work based on the same principles as everyday computers, but their performance levels are much higher and they look more like the classic mainframes of old. Unlike normal desktop PCs or laptops, they process massive data sets and perform calculations at incredible speeds. They're the fastest computers in the world, according to IBM, and require a massive amount of infrastructure to operate — including advanced cooling systems.
Architecturally, they are also fitted with many more components than normal PCs. Your laptop might have one central processing unit (CPU) and one graphics processing unit (GPU), but one supercomputer may have thousands upon thousands of CPUs and GPUs — each of them considerably more powerful than those you find off the shelf.
Their performance is also measured using floating-point operations per second (FLOPS) — where one floating-point operation is a mathematical calculation. The most powerful supercomputer in the world now exceeds 1 exaFLOP — 1 quintillion (1018) FLOPS — while normal PCs and laptops usually have power of several hundred gigaFLOPS — 1 trillion (109) FLOPS. We refer to machines like this as exascale supercomputers.
Because they can work with massive amounts of data and process calculations incredibly fast, scientists often use supercomputers to crack problems in drug and material discovery. Supercomputers can also make predictions — like forecasting the weather — and even learn how to play chess, like IBM's classic Deep Blue supercomputer in 1997.
Based on the latest TOP500 rankings, here are the seven most powerful supercomputers online today.
1. Frontier
- Location: Oak Ridge National Laboratory — Tennessee, U.S.
- Performance: 1,194 petaFLOPS (1.2 exaFLOPS)
- Components: AMD EPYC 64-core CPUs and AMD Instinct MI250X GPUs
- First online: August 2022
Currently top of the list, Frontier — built by supercomputing giant HPE Cray — became the first exascale computer in the world when it went online in 2022. Scientists initially planned to use Frontier for cancer research, drug discovery, nuclear fusion, exotic materials, designing superefficient engines and modeling stellar explosions, according to IEEE Spectrum.
In the coming years, scientists will use Frontier to design new transport and medicine technologies, reported MIT Technology Review. Evan Schneider, assistant professor in computational astrophysics at the University of Pittsburgh, told MIT Tech Review that he wants to run simulations of how the Milky Way has evolved over time.
2. Aurora
- Location: Argonne National Laboratory — Illinois, U.S.
- Performance: 585 petaFLOPS (0.59 exaFLOPS)
- Components: Intel Xeon Max Series CPUs and Intel Data Center Max Series GPUs
- First online: June 2023
One of the youngest supercomputers on the list may also one day become the most powerful. Based at the Argonne Leadership Computing Facility (ALCF), Aurora became the second exascale supercomputer ever, and ALCF representatives said it has the potential to reach 2 exaFLOPS of computing power — which is double Frontier's.
Built in partnership with Intel and HPE, Aurora integrates scientific tools and analysis, performs modeling and simulation and runs artificial intelligence (AI). Aurora's firepower allows it to create accurate models in various domains including climate projection, material science, energy storage and nuclear fusion. Fusion, in particular, is a focus for Aurora, according to HPCWire, and it may one day help to unlock the secret to large-scale fusion energy.
3. Eagle
- Location: Microsoft Azure — The Cloud / Distributed
- Performance: 561 petaFLOPS (0.56 exaFLOPS)
- Components: Intel Xeon Platinum 8480C 48C CPUs and Nvidia H100 GPUs
- First online: August 2023
Microsoft's Eagle supercomputer is isn't based in a laboratory — it is based in the cloud, and anybody can access it through the Microsoft Azure cloud platform. It's a distributed network of systems that collectively boast enough power to be the third fastest supercomputer in the world. Eagle can in theory be accessed by anybody willing to pay a fee.
4. Fugaku
- Location: Riken Center for Computational Science — Kobe, Japan
- Performance: 442 petaFLOPS (0.44 exaFLOPS)
- Components: A64FX CPUs
- First online: June 2020
Once the most powerful supercomputer in the world — between June 2020 and June 2022 — Fugaku is one of the oldest top-five systems on this list. It receives its name from Mount Fuji, an active volcano approximately 60 miles (100 kilometers) from Tokyo, and displaced the Summit supercomputer (No. 7 on the list) when it first reached the top of the TOP500 list.
Scientists have used Fugaku for several crucial research questions through the years. During the COVID-19 pandemic, researchers used its number-crunching abilities to confirm that face masks made with non-woven fabric were more effective at blocking airborne respiratory droplets, according to Nikkei Asia. Fugaku is currently training Japanese AI large language models, in the mold of ChatGPT, Japan News reported.
5. LUMI
- Location: CSC Data Center — Kajaani, Finland
- Performance: 380 petaFLOPS (0.38 exaFLOPS)
- Components: AMD 3rd-Gen EPYC 64-core CPUs and AMD Instinct MI250X GPUs
- First online: June 2021
LUMI, based in Finland, is Europe's most powerful supercomputer and the fifth-fastest in the world. It uses 100% renewable hydroelectric energy, according to European Union (EU) officials, and its waste heat is used to warm nearby buildings. It began running pilot operations three years ago and became fully operational in February 2023.
Designated as a supercomputer that researchers from across Europe can use for collaborative research, this system is optimized for AI-based workloads, officials have previously said. LUMI is also used as a "partner" for quantum computers, namely two systems called QAL 9000 and Helmi — both based in Finland. This quantum-classical computing partnership aims to give researchers the best of current quantum computers (which are severely limited based on their potential in the future) and supercomputers.
6. Leonardo
- Location: CINECA data center — Bologna, Italy
- Performance: 239 petaFLOPS (0.23 exaFLOPS)
- Components: Intel Xeon Platinum 8358 32-core CPUs and Nvidia A100 GPUs
- First online: November 2022
Another system that's part of the EU's EuroHPC supercomputing program, Leonardo comprises three modules that combine to make it Europe's second-fastest machine. The organization that hosts it, CINECA, is a consortium of Italian universities, public research centers and government agencies.
The supercomputer entered its pre-production phase in May 2023, during which time 80 projects were submitted and 13 chosen. Then, in August, the supercomputer entered its production phase.
7. Summit
- Location: Oak Ridge National Laboratory — Tennessee, U.S.
- Performance: 149 petaFLOPS (0.15 exaFLOPS)
- Components: IBM POWER9 22-core CPUs and Nvidia Tesla V100 GPUs
- First online: June 2018
Developed for use at the Oak Ridge National Laboratory, Summit was eight times more powerful than the institution's previous supercomputer, Titan, and was the most powerful supercomputer in the world for two years until it was displaced by Fugaku.
In addition to scientific modeling, Summit is optimized for AI workloads including machine learning and deep learning in areas such as human health and material discovery, representatives said in a statement.
The supercomputer has been pivotal for key research through the years, according to Oak Ridger. It was used, for example, to screen millions of drugs during the COVID-19 pandemic to determine if any could prevent the virus from entering or replicating in human cells. It's also been used to model turbulence.
