In late 2020, Los Alamos National Laboratory (LANL) — which operates under the Department of Energy’s National Nuclear Security Administration (NNSA) — completed installation of its new supercomputer, Chicoma (so named for a prominent mountain in New Mexico, and pictured in the header). Now, less than two years later, the lab is engaging Chicoma with a battery of 75 wide-ranging projects throughout 2022. The lab also disclosed new details on Venado, one of its next-generation systems.
Chicoma, a HPE Cray EX system networked with HPE Slingshot, boasts 688 nodes and more than 300TB of memory. 560 of the nodes are equipped with dual AMD Epyc 7H12 CPUs, while the remaining 128 are equipped with individual AMD Epyc 7713 CPUs and quadruple Nvidia A100 GPUs. HPCwire estimates that Chicoma offers more than eight petaflops of peak computing power. It is a somewhat unusual member of LANL’s computing arsenal; typically, the lab — which conducts highly classified weapons simulations for the NNSA — operates in relative secrecy, but Chicoma is one of five HPC systems available to LANL scientists and engineers for unclassified work. And beyond its availability, Chicoma also boasts some unique capabilities.
“Chicoma is the first system specifically designed to offer multiple architectures within a single large platform,” explained Carolyn Connor, interim program manager for institutional computing at LANL. “Chicoma’s flexible compute environment, including large-scale platform, architectural diversity and node capability, make it valuable for a wide range of scientific problems of interest to the Laboratory.”
The 75 projects for 2022 indeed span many disciplines, from astrophysics and climate science to nanotechnology and wildfires. LANL, though, is particularly highlighting the Covid research taking place on Chicoma. The lab said that Chicoma’s earliest users focused on producing meaningful Covid research out of the gate, which resulted in supercomputer-powered epidemiological modeling, advanced bioinformatics work and more. In this new research phase, the lab expects that focus to continue, including novel vaccine design work led by LANL biophysicist Gnana Gnanakaran that the lab says could help design vaccines that protect against diverse coronaviruses from SARS-CoV-1 to MERS.
The lab also discussed how Chicoma paves the way for Venado, its next institutional computing supercomputer (also named for a New Mexico peak). Venado, first announced 11 months ago, is the result of a codesign partnership between LANL, Nvidia and HPE scheduled for deployment to early users in 2023 with general availability in early 2024. When the system was announced, LANL revealed that it would use Nvidia’s next-generation, Arm-based Grace CPU; the new announcement confirms that it will also use Nvidia’s next-generation (“ANext”) A-series GPUs.
“Chicoma is a powerful resource and a stepping stone to future technologies aligned with our longer-term computing strategy,” said Irene Qualters, associate director for Simulation and Computing at LANL. “Chicoma provides Los Alamos scientists with state-of-the-art supercomputing capability and positions our researchers and applications to easily adopt the next generation of leading-edge computers.”
To learn more, read the reporting from LANL here.
This story has been updated with additional details.