Project Lead: Dr. Howard Shrobe
Sponsoring Organization: DARPA
Project Synopsis: SDH’s goal is to build runtime-reconfigurable hardware and software that enable performance close to that of application-specific integrated circuits (ASICs) – without sacrificing programmability for data-intensive algorithms, or incurring the cost, development time or single-application limitations associated with ASICs. The program defines data-intensive algorithms as machine learning and data science algorithms that process large volumes of data and use intense linear algebra, graph search operations and their associated data-transformation operators. SDH will create malleable hardware/software architectures that, unlike ASICs, allow an application to defer hardware configuration to runtime. By the end of the program, SDH expects its systems to be at efficiencies within 5X of ASICs and 500-1000X better than CPU implementations with the same programmability as current NumPy/Python implementations. The ultimate intent is to enable pervasive use of Big Data solutions in a wide range of DoD applications, including intelligence, surveillance and reconnaissance, predictive logistics, decision support and more.