U.S. should regain lead in supercomputing
Faster computers would bring big benefits in a host of endeavors
The United States no longer dominates the top positions on the list of the world’s fastest computers, as it once did. One look at the latest list tells the story: an American computer doesn’t even show up till fifth place. China holds the first two spots, followed by Switzerland and Japan.
But America can reclaim the top spots by continuing its strong commitments to partnerships in computing across industry, academia and federal research laboratories — and by continuing to commit funding for the next generation of supercomputers.
Energy Department scientists are now in a race to see who reaches the next big computing milestone, dubbed exascale computing, which means performing a quintillion calculations each second. That number is a 10 followed by 18 zeroes, and it’s roughly 10 to 20 times faster than the fastest computer today. There’s a lot at stake. Supercomputers are crucial to the science behind many of the high-tech tools, high-speed information systems, national security technologies, and medical advances that enrich and protect our lives every day.
The current fastest American computers push the frontiers of science and directly contribute to national security at the Energy Department national laboratories. These machines create detailed, realistic simulations of the complex systems and processes in some of the toughest challenges facing science. They include curing cancer and accurately modeling highly complex earth systems, such as the weather and ocean currents. They include explaining how the atoms in a material bond together and yield under pressure and tracing the myriad atomic-scale interactions of a nuclear detonation.
Consider the Trinity supercomputer at Los Alamos, the seventh-fastest in the world by one measure and third by another. Trinity is managed and operated by the Alliance for Computing at Extreme Scales (ACES) partnership of Los Alamos National Laboratory and Sandia National Laboratories and housed at Los Alamos. Geneticists rely on the computer to simulate the whirling dance of a billion atoms in the human genome. Theoretical biologists can use it to discover the genetic on/off switch for cancer. Medical researchers could consider new strategies for fighting genetic diseases like Down syndrome.
A billion atoms in a genome sounds like a lot, but the genome is made of thousands of times more atoms. To understand how the genome works at this most basic level, which will open whole new pathways for medicine, we need to follow each one of those atoms through a simulation. The current simulation brought Trinity to its knees. The next step will turn the lights out.
Scientists also use Trinity to simulate nuclear bomb detonations. Physicists combine theory with vast stores of data culled from actual nuclear tests completed years ago to create “movies” revealing the inner workings of the nuclear chain reaction of a bomb. This allows them to check the effectiveness of our current nuclear weapons without having to test them physically. Yet this complicated simulation takes Trinity to the very limits of its processing power and memory. It’s the same story with other Energy Department supercomputers.
What scientific breakthroughs lie beyond those limits? To find out, the Energy Department formed the Exascale Computing Project, a collaborative of Argonne, Lawrence Berkeley, Lawrence Livermore, Los Alamos, Sandia and Oak Ridge national laboratories. The first planned nextgen machines under the project will be developed and put to work at Argonne, Oak Ridge and Lawrence Livermore. With these superfast computers operating at exascale, scientists will solve problems that matter by simulating not just the individual parts of systems — current computers can do that — but whole systems in motion at an unprecedented level of detail.
That ability will open up previously unglimpsed horizons in science. As the next generation of supercomputers accelerates untold, stunning scientific advancements, scientists will begin solving problems they have not yet thought of. In national security, these problems might involve unanticipated questions about the aging nuclear stockpile or evaluating the threat posed by new, surprising technologies from our adversaries.
New solutions can improve the efficiency and stability of the electric power grid, uncover the secrets of earthquakes, and develop innovative new materials for safer, lighter, more fuel-efficient cars.
Science has served as a critical partner in advancing the U.S. economy for decades. Driven by the breakthroughs enabled by the next generation of supercomputers, whole new industries can emerge as entrepreneurs propel these innovations into the marketplace. New advances can also reinvigorate aging technologies.
America has long led the world in this kind of technology and innovation. By investing in the drive to build the world’s fastest supercomputer and achieving exascale, we can reassert our leading role in the global economy.