- Japan’s Fugaku supercomputer keeps its spot at #1
- Fugaku’s ARM-based architecture is unique for a top-performing supercomputer
- NVIDIA's Selene snags the #5 spot from Tianhe-2
We've updated this popular article to reflect the most current ranking.
Peak performance within supercomputing is a constantly moving target. In fact, a supercomputer is defined as being any machine “that performs at or near the currently highest operational rate.” The field is a continual battle to be the best. Those who achieve the top rank may only hang on to it for a fleeting moment.
Competition is what makes supercomputing so exciting, continually driving engineers to reach heights that were unimaginable only a few years ago. To celebrate this amazing technology, let’s take a look at the fastest computers as defined by computer ranking project TOP500—and at what these machines are used for.
5. Selene (US)
Selene has snagged the #5 spot from Tianhe-2. Selene’s performance has rocketed from a Linpack score of 27.5 petaFLOPS in the June 2020 TOP500 listing to 63.4 petaFLOPS in the current edition.
Coincidentally, Selene is also #5 on the GREEN500 list for its high energy efficiency.
Built by NVIDIA, this system came together during the early stages of the COVID-19 pandemic. “Skeleton crews” of engineers tediously followed safety protocols to finish the build. Created with artificial intelligence (AI) in mind, NVIDIA claims that Selene is capable of more than 1 exaFLOPS of AI performance.
4. Sunway TaihuLight (China)
A former number one, Sunway TaihuLight dominated the list for two years after its debut in June 2016. At that time, its 93.01 petaFLOPS and 10,649,000 cores made it the world’s most powerful supercomputer by a wide margin, boasting more than five times the processing power of its nearest competitor (ORNL’s Titan) and nearly 19 times more cores.
But given the non-stop pace of technological advancement, no position is ever secure for long. TaihuLight ceded the top spot to competitors in June 2018.
Located at the National Supercomputing Center in Wuxi, China, TaihuLight’s creators are using the supercomputer for tasks ranging from climate science to advanced manufacturing. It has also found success in marine forecasting, helping ships avoid rough seas and aiding offshore oil drilling.
3. Sierra (US)
Lawrence Livermore National Laboratory’s (LLNL) Sierra initially debuted at #3 on the June 2018 list with 71.6 petaFLOPS. Optimization later pushed the processing speed on its 1,572,480 cores to 94.6 petaFLOPS, earning it the #2 spot in November 2018. However, the ascension of a new number one in June 2020 pushed Sierra back to third position.
It has remained here for the November 2020 list.
Incorporating both IBM central processing units (CPUs) and NVIDIA graphics processing units (GPUs), Sierra is specifically designed for modeling and simulations essential for the US National Nuclear Security Administration.
2. Summit (US)
As part of the US Department of Energy’s renewed commitment to supercomputing power, Oak Ridge National Laboratory’s (ORNL) Summit first claimed the #1 spot in June 2018, taking the top rank from China for the first time in 6 years.
Since its debut on the June 2018 list, Summit improved its initial High Performance Linpack (HPL) performance from 122.3 to the current 148.6 petaFLOPS. Unusually for such a high-performing machine, Summit also initially ranked third on the GreenTop500, which measures energy efficiency in supercomputers, though it has since dropped to eighth place.
A seven-member team from ORNL won the 2018 Gordon Bell Prize for their deployment of Summit to process genetic data in order to better understand how individuals develop chronic pain and respond to opioids. Summit is now playing a critical role in the global race to discover treatments and vaccines against COVID-19.
1. Fugaku (Japan)
Jointly developed by RIKEN and Fujitsu, Japan’s Fugaku is the #1 fastest supercomputer in the world. Looking at this machine's specs, Fugaku's continued reign as champion isn’t surprising.
Boasting nearly 7.6 million cores and a speed of 442 petaFLOPS, Fugaku far outperforms Summit’s 148.6 PetaFLOPS.
Fugaku is the first top-ranked system to be powered by ARM processors. Two other new features are hybrid memory cubes attached to each of the processors, and a new iteration of the Tofu network that provides tight integration between all of the nodes in the system.
The ARM-based architecture represents a dramatic shift in the type of compute traditionally employed in supercomputers. Its designers view its success as proof that there is still room for innovation in HPC.
Installed at RIKEN Center for Computational Science (R-CCS) in Kobe, Japan, Fugaku is intended for applications that will address high-priority social and scientific issues. These include drug discovery, personalized medicine, weather and climate forecasting, clean energy development, and exploring the fundamental laws of the universe. It is already being used in an experimental basis for COVID-19 research. Fugaku will begin full operation in April 2021.
The race to possess the most powerful supercomputer never really ends. This friendly competition between countries has propelled a boom in processing power, and it doesn’t look like it’ll be slowing down anytime soon. With scientists using supercomputers for important projects such as curing debilitating diseases, we can only hope it will continue for years to come.
Read more about what's next for supercomputers, beyond petascale: