An approach to high-level synthesis using constraint logic programming

Presents a new method for modeling and solving high-level synthesis problems. In our approach, finite-domain constraints and the related constraint-solving techniques offered by constraint logic programming are used. They make it possible to define basic constraints on operations and registers and provide a way to find optimal or suboptimal solutions to the data-path synthesis problem. Different design styles, such as multi-cycling, chaining, pipelined components and algorithmic pipelining can be modeled in this framework. The proposed formulation combines different constraints in one representation, and thus the optimization can find a better solution. The prototype system has been implemented in the constraint logic programming system CHIP. The extensive experiments carried out using this system have proved the feasibility of the presented approach.