From a history-making quartet of lightbulbs powered by nuclear energy to discoveries enabled by the one of brightest light sources in the Western Hemisphere to insights into the dark corners of the universe, 75 years of Argonne research have produced breakthroughs that have changed our society and made our lives safer, healthier and more prosperous. This article is part of a 75th anniversary series describing Argonne’s history of discovery, current science program and future research thrusts.
U.S. Department of Energy’s (DOE) Argonne National Laboratory has deep roots in computational science and supercomputing. Argonne helped pioneer early computing research in 1953, when physicists at the laboratory built AVIDAC, the laboratory’s first digital computer. The size of a large room, AVIDAC could perform arithmetic operations for numbers up to 999,999,999,999. An arithmetic operation took 10 microseconds (there are 1 million microseconds per second), and a multiplication operation took 1 millisecond, or 1/300th the amount of time it takes to blink.
Today, Argonne’s supercomputers still fill large rooms, but they are immensely more powerful. Since being founded in 2006, the Argonne Leadership Computing Facility, a DOE Office of Science User Facility, has operated a series of increasingly powerful supercomputers dedicated to solving the nation’s most pressing challenges in science and engineering. Argonne is now readying for its next-generation supercomputer, Aurora, an exascale system scheduled to arrive in 2022 that will be capable of performing a billion billion (or million trillion) calculations per second.
To bolster research conducted on its supercomputing resources, Argonne is building technology around — and its reputation in — artificial intelligence (AI), which is playing an ever-increasing role in solving more complex and data-heavy problems. Researchers across the laboratory complex are using an array of AI techniques to design better materials and processes, safeguard the nation’s power grid, accelerate medical treatments and automate traditional research. Additionally, Aurora is being designed to handle large-scale AI and data analysis workloads alongside traditional modeling and simulation campaigns, providing researchers with a suite of powerful tools for accelerating scientific discoveries.