Lawrence Livermore National Laboratory: programs, progress and news

Analyst view

Operational November 2024 with 1.809 exaflops. Ranked 2nd on June 2026 TOP500.
Built by HPE with AMD EPYC CPUs and Instinct MI300A accelerators; 100% fanless direct liquid cooling.
Primary mission: Stockpile Stewardship Program ensuring U.S. nuclear arsenal safety and effectiveness.

As reported

Processor Architecture

4th generation AMD EPYC Zen 4 processors paired with Instinct MI300A GPUs; 11M cores, 44,000+ APUs.

Cooling Technology

HPE direct liquid cooling eliminates fans, improving efficiency and reliability.

Software Stack

ROCm enables efficient utilization of the hybrid CPU-GPU architecture.

Programs

Lawrence Livermore National Laboratory, two programs

Updated 1 Jun 2025

Research · Proto · Pilot · Deploy · Scaling

Open · Active · Partial

Maintaining stockpile simulation accuracyActive research
Aging weapons require higher-fidelity simulations without underground testing.
Programming efficiency on heterogeneous systemsActive research
Optimizing codes for both CPUs and GPUs remains an active challenge.

Company-wide

2032EYC laser upgrades toward 2.6 MJ capacityNational Ignition Facilityby ~75 mo

AI factories & supercomputers · 8

1 customer · 2 suppliers · 2 partners

Customer
National Nuclear Security Administration (opens energy.gov)
NNSA funds and directs LLNL's mission priorities, including nuclear security and stockpile stewardship.
Supplier
Hewlett Packard Enterprise (opens hpe.com)
HPE built and delivered the El Capitan system using its Cray EX255a architecture.
Supplier
Advanced Micro Devices (opens amd.com)
AMD provides processors, GPUs, and software platform for El Capitan.
Partner
Los Alamos National Laboratory (opens lanl.gov)
Fellow NNSA lab; El Capitan serves the Tri-Lab partnership (LLNL, LANL, Sandia) for stockpile stewardship.
Partner
Sandia National Laboratories (opens sandia.gov)
Fellow NNSA lab; shares access to El Capitan computing resources.

Key people