Most of LANL's production supercomputing now runs on open-source OpenCHAMI
Seven Lab clusters with more than 11,000 compute nodes have moved to the systems-management framework Los Alamos helped found in 2023.
LOS ALAMOS — Los Alamos National Laboratory has moved most of its production computing onto OpenCHAMI, an open-source framework for managing high-performance computing systems, the Lab announced Nov. 12. Seven Lab production clusters with more than 11,000 compute nodes now run on it, the release says.
OpenCHAMI grew out of an open community collaboration launched in 2023. Its founding members were Los Alamos, the National Energy Research Scientific Computing Center (NERSC) at Lawrence Berkeley National Laboratory, the Swiss National Supercomputing Center, HPE and the University of Bristol; Dell joined in 2024.
According to the release, computing centers have tended to install and configure systems and then resist change, with upgrades complicated by legacy tools. OpenCHAMI is cloud-native, lets a site deploy its own preferred tools and works with both HPE and Dell hardware, the Lab says.
Markus Berndt, the Lab's High Performance Computing division leader, said, “we've already seen reduced downtime for upgrades.” The Lab says the flexibility also helps its modeling and simulation and artificial intelligence systems (the Lab's term is “super intelligence”), for example by quickly changing the operating system on a subset of nodes.