Unifying behavioral synthesis and physical design

eously demand shorter and less costly design cycles. Designing at higher levels of abstraction makes both objectives achievable, but enabling techniques like behavioral synthesis have met high resistance in commodity markets. The reasons behind the resistance are twofold. First, such tools are unable to handle designs of the size that require designers to work at the behavioral level, and second, they have not been able to accurately incorporate physical characteristics or controller information into high-level decisions. Ignoring such considerations in the deep submicron era leads to frequent time-consuming iterations between behavioral and physical tools. Objective: The goal of the proposed work is to create algorithms and a methodology unifying behavioral synthesis and physical design processes in a manner not compromising transformations in either domain. The approach will be efficient enough to handle designs orders of magnitude larger than are currently feasible, providing an exp...

• Type: Conference
• Year: 2000
• Where: DAC
• Authors: William E. Dougherty, Donald E. Thomas
• Keywords: Accurate Physical Estimations, Behavioral Synthesis, Behavioral Synthesis Capabilities, DAC 2000, Design Automation