Heuristic search based approach to scheduling, allocation and binding in Data Path Synthesis

In this paper, we present a novel heuristic search based approach (VITAL) to Data Path Synthesis. The approach is flexible as it is based on a generalized heuristic search algorithm and is capable of trading off between the solution quality and the computation time. The search algorithms are based on computation of cost estimates which are lower bound on the achievable cost. The approach achieves a higher degree of integration compared to other approaches, and provides closer cost estimates by performing allocation and partial binding sub-tasks during scheduling. It supports a variety of design styles, design constraints and component types, resulting in a versatile system. Experiments on several existing benchmark designs show that even VITAL-NS, which follows strictly one path in the search tree, gives excellent results