Contributor

The next-generation high performance computing (HPC) environment is heading towards new levels of complexity to meet the growing user demand from HPC workflows. These workflows aim to integrate simulations and modeling, artificial intelligence (AI), and experimental and observational data to enable new modes of scientific discovery. Workflows typically span beyond a single HPC system and even beyond a single HPC facility, moving toward a multi-site Integrated Research Infrastructure (IRI). This talk describes the growing challenge, how it impacted our procurement and benchmarking approaches to prepare for the NERSC-10 system at the National Energy Scientific Research Computing Center (NERSC), and how it may shape our testing strategies.