China holds the top two spots for fastest computers in the world, and Switzerland holds the third, with the U.S. in the fourth, fifth and sixth spots.
But the U.S. might not miss its top spot for long. The Department of Energy awarded six companies a total of $258 million last Thursday to further the research and development of the world’s first exascale supercomputer. There are no computers that powerful today.
The U.S. formerly held the third spot, but this time it was edged out by a system from the Swiss National Supercomputing Centre, which moved up from eighth place. This is only the second time in 24 years of compiling the Top500 list that the U.S. did not have a computer place in one of the top three positions.
These computers process at petascale speeds, meaning their capabilities are measured in terms of one quadrillion (1,000,000,000,000,000) calculations per second. To put that in perspective, consumer laptops now operate at gigascale, which is one billion calculations per second.
The U.S. companies that received government funding — Hewlett Packard, Intel, Nvidia, Advanced Micro Devices and Cray — will all work to solve problems in energy efficiency, reliability and overall performance of a national exascale computer system.
An exascale computer is capable of processing a quintillion (1,000,000,000,000,000,000) calculations per second. That’s about a trillion times more powerful than a consumer laptop.
Exascale-level computing would allow scientists to make extremely precise digital simulations of biological systems, which could uncover answers to pressing questions like climate change and growing food that can withstand drought.
“As you develop models that are more sophisticated that include more of the physics, chemistry and environmental issues that are important in predicting the climate, the computing resources you need increases,” said Thom Dunning, a chemistry professor at the University of Washington and the co-director of the Northwest Institute for Advanced Computing.
Chemists are leading a lot of the advances in computing power, since advanced biological modeling requires really powerful processing. With more detailed biological modeling, chemists can, for example, learn how plant cells react to drought, which can help to better engineer crops — a project Dunning is working on with his research group.
The more powerful the computer, the more realistic the models are, which in turn provide scientists with more reliable predictions about the future and more concrete recommendations about what companies and governments need to do.
Exascale computing would also have a tremendous impact on the country’s national security. The National Security Agency and other law enforcement organizations collect more data in their dragnet digital surveillance operations than can often be processed in a timely, meaningful way, according to Dunning. With higher processing power, that data can be analyzed quickly to assess and predict potential threats.
The companies awarded the grants will cover at least 40 percent of the cost of the research projects themselves.
“Creating an exascale computer is well beyond anything that a private company can do on its own,” said Dunning, who added that building an exascale computer is a multibillion dollar effort.
U.S. investment in building an exascale machine will have benefits beyond just finishing the computer itself. The research and development gleaned along the way will flow down into lower-level systems that will give the U.S. a competitive advantage in terms of making powerful computing much more affordable and accessible, Dunning said.
Here’s a list of the Top 10 most powerful supercomputers in the world. The U.S. holds the most spots on the list, with five supercomputers that made the cut.
- Sunway TaihuLight — China
- Tianhe-2 (MilkyWay-2) — China
- Piz Daint — Switzerland
- Titan — United States
- Sequoia — United States
- Cori — United States
- Oakforest-PACS — Japan
- K Computer — Japan
- Mira — United States
- Trinity — United States
This article originally appeared on Recode.net.