A design methodology for interactive behavioral synthesis
- Author(s): Gajski, Daniel D.
- Juan, Hsiao-Ping
- et al.
Due to the recent increases in design complexity, behavioral synthesis has become an important area of research and company interest. However, there has been market resistance to accepting the automatic behavioral synthesis approach as a practical solution in general because, first, it often produces results inferior to manual designs, and second, it allows only minimum user control. To develop a feasible approach for behavioral synthesis to overcome the hurdles faced by the automatic approach, we propose interactive behavioral synthesis, which attempts to maximally utilize the human designer's insights. Using interactive behavioral synthesis, the users can control the design process, observe the effects of design decisions, and manually override synthesis algorithms at will. In this report, we present a design methodology as how the user interacts with an interactive behavioral synthesis system, which in contrast to an automatic synthesis system, enables the human designer fine-grain control over each synthesis task and continually supplies feedback in the form of quality measures so that the user can make informed design-related decisions. To demonstrate the proposed design methodology, we also present in this report a walk-through square-root approximation (SRA) example.