A Scheduling Framework for System-level Estimation

This paper presents the first main task of a new system estimation framework for codesign located before the partitioning and architectural estimation steps. The key issue of the proposed method is a dynamic estimation at an abstract level (without architectural target) based on a global estimation algorithm. The proposed algorithm, addresses a given function extracted from a complex application. It is based on function orientation and combines various efficient methods to quickly explore the design space and provides the next codesign steps with a small set of candidate solutions.