Skip to main content
A team led by Oak Ridge National Laboratory researchers used Frontier to explore training strategies for one of the largest artificial intelligence models to date. Credit: Getty Images

A team led by researchers at ORNL explored training strategies for one of the largest artificial intelligence models to date with help from the world’s fastest supercomputer. The findings could help guide training for a new generation of AI models for scientific research.
 

Frontier supercomputer sets new standard in molecular simulation

When scientists pushed the world’s fastest supercomputer to its limits, they found those limits stretched beyond even their biggest expectations. In the latest milestone, a team of engineers and scientists used Frontier to simulate a system of nearly half a trillion atoms — the largest system ever modeled and more than 400 times the size of the closest competition.

From left, J.D. Rice, Trevor Michelson and Chris Seck look at a monitor in Seck’s lab. The three are wearing safety glasses to protect against the laser beams used by the scanning vibrometer, which is helping Seck quantify vibration of an appliance in his lab. Carlos Jones/ORNL, U.S. Dept. of Energy

ORNL scientists are working on a project to engineer and develop a cryogenic ion trap apparatus to simulate quantum spin liquids, a key research area in materials science and neutron scattering studies.

ORNL researcher Felicia Gilliland loads experiment samples into position for the newly installed UR5E robotic arm at the BIO-SANS instrument. The industrial-grade robot changes samples automatically, reducing the need for human assistance and improving sample throughput. Credit: Jeremy Rumsey/ORNL, U.S. Dept. of Energy

The BIO-SANS instrument, located at Oak Ridge National Laboratory’s High Flux Isotope Reactor, is the latest neutron scattering instrument to be retrofitted with state-of-the-art robotics and custom software. The sophisticated upgrade quadruples the number of samples the instrument can measure automatically and significantly reduces the need for human assistance.

A newly completed tunnel section will provide the turning and connecting point for the Spallation Neutron Source particle accelerator and the planned Second Target Station. Credit: ORNL, U.S. Dept. of Energy

The new section of tunnel will provide the turning and connecting point for the accelerator beamline between the existing particle accelerator at ORNL’s Spallation Neutron Source and the planned Second Target Station, or STS. When complete, the PPU project will increase accelerator power up to 2.8 megawatts from its current record-breaking 1.7 megawatts of beam power.

Quietly making noise: Measuring differential privacy could balance meaningful analytics and identity protection

To balance personal safety and research innovation, researchers at ORNL are employing a mathematical technique known as differential privacy to provide data privacy guarantees.

The AI for Energy Report provides a framework for using AI to accelerate decarbonization of the U.S. economy. Credit: Argonne National Laboratory

Groundbreaking report provides ambitious framework for accelerating clean energy deployment while minimizing risks and costs in the face of climate change.

The transportation and industrial sectors together account for more than 50% of the country’s carbon footprint. Defossilization could help reduce new emissions from these and other difficult-to-electrify segments of the U.S. economy.

Scientists at Oak Ridge National Laboratory and six other Department of Energy national laboratories have developed a United States-based perspective for achieving net-zero carbon emissions. 

The EPA approved the registration and use of a renewable gasoline blendstock developed by Vertimass LLC and Oak Ridge National Laboratory that can significantly reduce vehicle emissions when added to conventional fuels. Credit: Adam Malin/ORNL, U.S. Dept. of Energy

The U.S. Environmental Protection Agency has approved the registration and use of a renewable gasoline blendstock developed by Vertimass LLC and ORNL that can significantly reduce the emissions profile of vehicles when added to conventional fuels. 

colors

Simulations performed on the Summit supercomputer at ORNL are cutting through that time and expense by helping researchers digitally customize the ideal alloy.