Part 1 - Constraint-driven Physical Design Speeds Custom IC Design Convergence

In this introductory Part I of V of this blog I will discuss the advanced node design challenges impacting CIC design convergence and the solutions to achieve expedited physical implementation convergence.

As designers move to 65nm technologies and below, the convergence of performance-driven design constraints and yield-driven manufacturing constraints intensifies the demand for new approaches for integrated circuit (IC) physical design (Figure 1).

At these geometries, more complex manufacturing effects dramatically impact the way engineers need to tune physical designs for optimal performance and yield. Besides addressing familiar speed and capacity concerns, advanced physical design requires an architectural approach that emphasizes quality of results, more effective convergence across a broader array of constraints, and significantly greater control by designers of the physical design process itself.

Faster routing of larger designs is not enough, and design convergence means much more than area, timing and power. Instead, designers need the ability to analyze routing more effectively, and incrementally improve both performance and yield with each physical design iteration.

As the electronics industry continues to drive toward more advanced manufacturing technologies, semiconductor companies face shrinking product lifecycles and rising demand for greater functionality. For engineers, each advance in design and manufacturing capabilities brings greater challenges in every phase of development, yet dictates a greater need to reach closure on a growing list of divergent constraints arising from each stage in the development cycle.

As engineering teams move designs from high-level and detailed logic design to floorplanning and routing in physical design, they must work collaboratively to ensure that physical design maintains tight objectives for design performance, functionality and manufacturability. Accordingly, physical design and verification needs to work smoothly in the design flow, efficiently providing detailed results needed to ensure high quality results within tightening product schedules.

Yet, as designs move to deep nanometer technologies at 65nm and below, designers find that electronic design automation (EDA) tools developed even for 90nm designs are unable to address the further challenges associated with these advanced technologies. Inevitably, the lack of precise analysis of device performance at these new geometries forces design teams to make tradeoffs and concessions to ensure manufacturability.

In Part II thru V of this blog I will discuss the sub-topics of Growing Complexity, Successive Refinement, Constraint-driven Physical Design, and the Fundamental Change required for advanced node CIC design convergence.

Craig Thompson