Hewlett Packard Enterprise (HPE) this morning said it has won a $35+ million contract to build a supercomputer for the National Center for Atmospheric Research (NCAR), a federal geoscience R&D center for meteorology, climate change and solar activity.
HPE said the CPU/GPU-powered system, funded by the National Science Foundation, is expected to deliver 3.5x the power of NCAR’s existing HPC system, is slated for deliver later this year to the NCAR Wyoming Supercomputing Center (NWSC) in Cheyenne and put to use in 2022.
The system, known for now as NWSC-3, is expected to rank among the world’s 25 fastest supercomputers. HPE said it will integrate the following HPC technologies within its HPE Cray EX supercomputer architecture to enable a theoretical peak performance of 19.87 petaflops:
- 60 petabytes of storage to support complex workloads in modeling, simulation and AI using the Cray ClusterStor E1000
- 692 terabytes of total memory
- HPE Slingshot (v11) interconnect in a Dragonfly topology, with 200 Gb/sec bandwidth per port per direction, for congestion control
- An integrated software suite for HPC and AI applications using the HPE Cray Programming Environment
- 2,570 compute nodes: 2,488 homogeneous compute and 82 heterogeneous (GPU) nodes; the system will be powered by 2nd and 3rd Gen AMD EPYC CPUs and Nvidia A100 Tensor Core GPUs; the system is expected to get 20 percent of its sustained computing capability from GPUs, according to NWSC, which said it will consume 40 percent more electrical power than Cheyenne
- HPE Cray Operating System, a tuned version of SUSE Linux; Altair Accelerator Plus scheduler with PBS Professional Workload Manager; support for Docker and Singularity containers, and containers that support the Open Container Initiative standard; HPE Cray Programming Environment, support for OpenMP 4.5 and 5.0, and MPI v3.1; Nvidia AI and HPC software with AI frameworks, compilers, libraries, models and tools
Of the work to be done with the new system, Anke Kamrath, director, NCAR Computational and Information Systems Laboratory, said, “The resulting research will lead to new insights into potential threats ranging from severe weather and solar storms to climate change, helping to advance the knowledge needed for improved predictions that will strengthen society’s resilience to potential disasters.”
While NCAR’s existing system is known as Cheyenne, the center is holding a contest for Wyoming schoolchildren to name the new system. NCAR said research targets will include:
- Improving predictions of seasonal water supply, drought risk and flooding through detailed modeling and forecasting tools to inform water management experts, public utilities and farmers
- Managing wildfire risk by simulating representations of physical processes in a region to help forecasts in wooded areas becoming prone to fires. Simulations will factor in data from local winds and air density, soil moisture, and vegetation patterns such as types of grass and leaves.
- Foreseeing hazards and impacts of climate change from extreme weather conditions, such as thunderstorms, tornadoes and hurricanes.
- Understanding the dangers of solar storms using three-dimensional simulations of the sun’s turbulent plasma flows and magnetic fields to enable predictions of potential solar impact that can disrupt the earth’s atmosphere and trigger space weather events that threaten communications systems and power grids.
The NWSC said that since its opening in 2012, more than 4,000 users from nearly 625 universities and other institutions in the U.S. and other countries have used its resources. Last year, the center joined the COVID-19 High Performance Computing Consortium for research into the coronavirus.
“We are inspired by leading R&D centers, such as (NCAR), in their research efforts of applying HPC to understanding earth systems, including severe weather events and wildfires, that lead to informed decisions in earth’s keeping people safe and preserving ecosystems,” said HPE’s Bill Mannel, VP/GM, HPC. “We are honored to have been selected by NCAR to build its next-generation supercomputer that will tackle complex computational research for a range of geosciences while also making it easier for researchers to store and make that data accessible to help communities make smarter decisions.”