AMD-powered Lux supercomputer opens to researchers at Oak Ridge
Reported by letsdatascience, StreetInsider, Securities.io and 3 more outlets6 sources ↓
Oak Ridge National Laboratory has opened the Lux supercomputer to scientific users under a public-private partnership in which AMD is co-investing. The system will support research combining AI with high-performance computing.
Key points
- Lux entered the TOP500 ranking at No. 14
- The system has more than 4,000 AMD Instinct MI355X GPUs across more than 500 nodes
- Lux offers 3.5 million node-hours annually, with half reserved for Genesis Mission research
- Initial projects include fusion plant design, materials development and Earth-system prediction
Sources
- letsdatascienceRead ↗HPE and AMD Lux Supercomputer Opens at Oak Ridge, Ranks 14th
- StreetInsiderRead ↗AMD-powered Lux supercomputer opens to researchers at Oak Ridge
- Securities.ioRead ↗AMD and DOE Open Co-Invested Lux Supercomputer to Researchers
- TechPowerUpRead ↗AMD-Powered Lux Supercomputer Opens to Scientific Users at Oak Ridge
- Business InsiderRead ↗AMD-Powered Lux Supercomputer Opens to Scientific Users at Oak Ridge, Becoming First AI Factory for DOE’s Genesis Mission
- finance.biggoRead ↗AMD Powers 202 of World's Fastest Supercomputers as DOE's Lux Debuts at No. 14
Details
HPE built Lux for the U.S. Department of Energy’s Oak Ridge National Laboratory using AMD EPYC processors, Instinct MI355X GPUs and Pensando networking. The system has more than 4,000 GPUs across more than 500 nodes, with 1.1 petabytes of high-bandwidth memory in total. The TOP500 list released alongside the opening ranked Lux No. 14.
Lux will provide 3.5 million node-hours each year. Half are dedicated to Genesis Mission research; the remaining half are reserved for paid access by public- and private-sector partners, according to Securities.io. AMD is co-investing in the system and plans to use some reserved time for development and testing while making part available to partners.
Projects allocated to Lux include AI-assisted fusion power plant design, materials development and advanced manufacturing, and Earth-system prediction focused on water availability, streamflow and river temperature in the Tennessee Valley Authority region. Securities.io reports that Lux supports 40 exaFLOPS for AI inference using FP4, 20 exaFLOPS for AI training using FP8 and 317 petaFLOPS for FP64 modeling and simulation; these figures use different workloads and precisions.
Written from coverage as of
Images
- TechPowerUp
- finance.biggo
- Securities.io
- finance.biggo
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh