Flexible controlpath microarchitecture synthesis based on artificial intelligence

The integration of methods from artificial intelligence in the synthesis of microarchitectures for processor control paths is described. A general synthesis method requires exploration of the design space for optimal microarchitectures. With the introduction of a design-state abstraction for a microarchitecture, the synthesis problem can be formulated as a search problem in a design-state space. In this space, existing synthesis algorithms act as design-actions and perform local design-state transitions. Artificial intelligence provides the global synthesis process with tools for state space search and action planning. The prototype system demonstrates the power and flexibility of this approach