Incorporating New Technologies to Meet the Demands of Next Generation Problems

HPC applications have diverse requirements and solve a variety of problems. Some need a few to hundreds of thousands of cores to complete independent tasks. Others divide and synchronize work across tightly connected infrastructure. Some workflows or ensembles require different parameters, decompositions, or phases as part of problem solving, and can benefit from running on a variety of architecture types. As HPC applications, ensembles, and workflows adapt to solve the next set of problems, and the community looks to bridge the gap between HPC in the cloud and cloud technologies in HPC, we discuss ways to identify and embrace the benefits of each to improve efficiency, sustainability, and performance.