SHF: Small: Parallel Algorithms and Architectures Enabling Extreme-scale Graph Analytics for Biocomputing Applications

Washington State University, Pullman WA

Investigators

Abstract

Graph-theoretic modeling of biological data has a rich history of delivering foundational scientific knowledge and breakthrough discoveries. As data sets continue to explode both in size and complexity, the combination of graph analytics and scalable (parallel) computing has a critical role to play in shaping the future of data-driven discovery in many biological applications including national health. Yet, implementing such graph computations at scale continues to be a daunting challenge despite the growing availability of high-end parallel architectures. The goal of this project is to design efficient parallel algorithms and architectures that would enable extreme scaling of graph computations in biological applications. Other project activities integrate and leverage upon the research outcomes of this project, while preparing the next generation scientific workforce. The project is also leading to the development of curricular modules in parallel algorithms and applications, and related hardware design, and conference tutorials for broader outreach. The project is focused on developing core techniques in two problem spaces: i) performing graph analytics at scale for a host of generic graph operations that find prevalent use-cases in biological applications and also in many other data-driven domains; and ii) performing graph construction at scale using biological raw data. Taken together, the proposed effort embodies a systematic and holistic approach to enhance the reach and impact of parallel computing on large-scale graph applications and, in the process, usher in new generic data-driven design techniques and paradigms into parallel applications design. While the emphasis will be on biological applications, as a space for drawing scientific motivation and to demonstrate utility through validation and testing, it is expected that many of the developed techniques will extend beyond this realm and impact a broader class of applications that need extreme-scale processing of graphs. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.