Department of Energy invests $200 million for Argonne supercomputer

The Argonne National Laboratory will collaborates with Intel and Cray to develop a supercomputer using a $200 million grant from the Department of Energy.

By Stephanie Williams

On Wednesday, April 8, the U.S. Department of Energy (DOE) announced plans to invest $200 million in the supercomputing efforts of Argonne National Laboratory. The investment was one of two High Performance Computing (HPC) grants awarded by the DOE, as part of an effort to advance U.S. leadership in large-scale computing. Argonne National Lab is managed by UChicago Argonne, LCC, and Jacobs Engineering Group Inc., and will be responsible for the new supercomputer project, Aurora.

Argonne National Laboratory is managed by UChicago Argonne, LLC, for the U.S. Department of Energy’s Office of Science and will be responsible for building the new supercomputer, Aurora.

Argonne will collaborate with Intel and Cray to develop a next-generation supercomputer system, which is slated for completion in 2018. The supercomputer system will be used by researchers in scientific and engineering fields to assist in research on materials science (batteries and solar panels), biological science (biofuels and disease control), transportation efficiency (aerodynamics and efficient engines), and renewable energy (wind turbine design).

“Aurora will deliver the system capabilities that our users require to expand their investigations both in scale and scope,” Michael E. Papka, deputy associate laboratory director for computing, environment and life sciences at Argonne Leadership Computing Facility, said in a statement.

Aurora will have more advanced supercomputing abilities than Argonne’s current project, Mira. It will exceed Mira’s computational performance by a factor of eighteen, said Susan Coghlan, deputy division director of Argonne Leadership Computing Facility, in an Ask Me Anything (AMA) online discussion on Friday.

Coghlan compared the airflow modeling capabilities of Mira and Aurora to illustrate the differences between the two computing systems. While Mira can simulate the flow of air over a small part of wind turbine blades, Aurora can simulate flow over the entire blade. This difference in simulation ability will allow for nuance: it will allow scientists to “investigate how vibrations from one blade impact that of other blades,” Coghlan stated.

The AMAs held by Coghlan and Papka on Friday highlighted public interest in Aurora’s impact outside of the computer lab. Coghlan described the widespread nature of the Aurora system’s predicted impact and enumerated a few areas that impact everyday life; Aurora’s computing power will allow researchers to “design better batteries, model weather at a higher resolution, and understan[d] the origins of the universe.”

To follow the supercomputer’s progress, directors Papka and Coghlen recommend checking Aurora’s website.