SciDAC Institute for Computer Science, Data, and Artificial Intelligence
The Department of Energy (DOE) has led the world in enabling large-scale high-performance computing as an indispensable tool for scientific discovery in a broad range of disciplines. Our newest generation of systems is not simply larger than past systems – it brings distinct and novel challenges. These challenges arise from the deepening of the memory hierarchy and inclusion of fast, nonvolatile storage within the HPC system; the widening ratio of peak FLOPS to I/O bandwidth; and the increasing ubiquity of computation accelerators. Moreover, the breadth of science performed on DOE advanced computing resources is growing and new motifs of investigation are emerging, including those involving experimental and observational data and artificial intelligence (AI). To address these challenges, we propose the RAPIDS2 SciDAC Institute for Computer Science, Data, and Artificial Intelligence.
The objective of RAPIDS2 is to assist the Office of Science (SC) application teams in overcoming computer science, data, and AI challenges in the use of DOE supercomputing resources to achieve scientific breakthroughs. To accomplish this objective, the Institute will:
• Engage with SciDAC Partnerships to identify technical challenges and potential solutions
• Solve computer science, data, and AI technical challenges for SciDAC and SC science teams,
• Work directly with SC scientists and DOE facilities to deploy and support our technologies, and
• Coordinate with other DOE computer science applied mathematics activities, and ECP, to maximize
RAPIDS2 will build on the successes of the RAPIDS Institute and will be organized around five primary thrust areas. The cornerstone of the project is application engagement and community outreach (§2.1). A rigorous plan for engagement through direct interactions, with existing and also prospective SciDAC Partnership teams, as well as outreach to connect with the broader community, will allow us to have the greatest and most immediate impact.