INDEPENDENT PUBLICATION · AMERICAN AMBITION BECOME A FOUNDING SPONSOR NEWS & ANALYSIS / DATED AT PUBLICATION

COMPUTING / LATEST COVERAGE

More computing time.
More room for science.

Oak Ridge National Laboratory has opened Lux to scientific users. The Tennessee supercomputer gives researchers a new way to pair Super Intelligence (SI) with large simulations.

A machine scientists can use.

Sometimes the most useful news about a supercomputer is that researchers can finally get time on it. On November 16, the lab announced that its Lux system was open to scientific users. It named fusion energy, advanced materials and manufacturing among the projects allocated time on the new machine.

The facility’s system guide identifies Lux as the first major addition to the U.S. Department of Energy’s computing infrastructure in support of the Genesis Mission. The department and chipmaker AMD jointly invested in the system, which uses hardware from Hewlett Packard Enterprise.

From simulation to training.

One useful feature is the connection between Lux and Frontier, Oak Ridge’s larger supercomputer. The facility says researchers with the right allocations can run a simulation on Frontier, write its output to the shared Orion storage system, and then train a model on that data using Lux.

One project allocated time on Lux plans to forecast streamflow, river temperature and water availability in the Tennessee Valley Authority region. Better forecasts could help operators plan when they have water for hydropower or for cooling thermal power plants. That is a research goal, not a result the team has already reported.

How the time is divided.

Oak Ridge’s opening announcement puts the system’s annual capacity at 3.5 million node-hours, with approximately half for the Genesis Mission and the rest available to AMD. A node-hour is one hour of use of one computing node. It is a measure of allocated computing time, not a count of researchers or an estimate of discoveries.

The opening gives research teams something practical: computing time to use simulation data, train models and decide what to test next. Our earlier coverage of the Genesis Mission explains the broader effort to get more computing resources into scientists’ hands.

Keep exploring

The infrastructure behind the work.

Explore American infrastructure ↗