In a collaborative effort aimed at advancing national security and fundamental scientific missions, Los Alamos National Laboratory, in partnership with Hewlett Packard Enterprise (HPE) and NVIDIA, has worked together to design and install Venado, the lab’s latest supercomputer.
On Monday, April 15, representatives from the U.S. Department of Energy, the National Nuclear Security Administration, Los Alamos, HPE, and NVIDIA gathered to unveil Venado. This new system features the HPE Cray EX supercomputer with NVIDIA GH200 Grace Hopper Superchips, at the Laboratory’s Nicholas C. Metropolis Center for Modeling and Simulation.
Thom Mason, director of Los Alamos National Laboratory, highlighted how Venado contributes to their state-of-the-art supercomputing capabilities, which drive advancements in national security, basic research, and the integration of artificial intelligence to tackle emerging challenges.
The NVIDIA Grace Hopper Superchips integrate an Arm-based NVIDIA central processing unit with an NVIDIA Hopper architecture-based graphics processing unit to address high-performance computing and large-scale artificial intelligence applications. These superchips can execute millions more instructions per second, typically at a lower cost and power consumption compared to previous chip technologies. Early testing of Venado has shown promising outcomes in atomistic simulations for materials science and high-resolution astrophysics simulations.
David Turk, deputy secretary of the U.S. Department of Energy, emphasized the critical role of supercomputing capabilities in addressing key challenges faced by national laboratories. He expressed anticipation for the impact of the Venado system in delivering new and significant results in various areas of interest by incorporating artificial intelligence methodologies.
Venado’s computing capacity will house 2,560 direct, liquid-cooled Grace Hopper Superchips within the exascale-class HPE Cray EX supercomputer. Additionally, the system will utilize 920 NVIDIA Grace CPU Superchips, marking it as the first large-scale system in the United States with this technology. NVIDIA has integrated 144 Arm cores in a Grace CPU Superchip to provide an immediate performance enhancement across a wide array of HPC applications.
Ian Buck, vice president of hyperscale and HPC at NVIDIA, praised Venado as the first U.S. supercomputer powered by NVIDIA Grace Hopper, delivering exceptional performance and energy efficiency to accelerate scientific breakthroughs in fields such as materials science, renewable energy, and astrophysics.
The Venado supercomputer, named after a peak in New Mexico’s Sangre de Cristo Mountains, was designed to enhance overall performance and workflow efficiency. It is connected with HPE Slingshot 11 interconnect to boost the system’s performance. Furthermore, additional HPE Cray supercomputing software enhances modeling and simulation workloads, aiding researchers in making new discoveries through AI and machine learning.
Trish Damkroger, senior vice president and general manager of HPC & AI Infrastructure Solutions at HPE, emphasized that the powerful supercomputing capabilities with AI-driven features will empower researchers to achieve significant breakthroughs and address real-world challenges in the evolving era of scientific discovery. The ongoing collaboration between Los Alamos, HPE, and NVIDIA showcases continuous innovation in supercomputing that accelerates discoveries in materials science and energy research.
Venado is the result of a co-design process that lays the foundation for ongoing collaborations focused on developing a wide range of computing, memory, and software technologies. This collaborative effort involves vendors, hardware architects, system software developers, domain scientists, computer scientists, and applied mathematicians working together to make informed decisions regarding hardware and software components.
