Circuit Placement Automation with Structured Regularity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Integrated circuit design faces challenges in structured placement optimization due to manual effort requirements, loss of regularity in standard cell random placement, and the computationally intractable problem of finding ideal orientations for cells, leading to suboptimal circuit designs with increased power, area, and congestion.
Innovation Solution
A method for generating placed, routed, and optimized circuit designs that utilize relative placement rules to maintain structured placement, allowing for the movement and optimization of circuit elements while adhering to specific positioning rules, and incorporating both constrained and unconstrained elements for improved design utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If standard cell random placement is used, then automation is achieved, but regularity and structured placement requirements are lost
Solution Approach 1:
The patent segments the placement process into two distinct phases: structured placement (which establishes regularity and organization) followed by standard cell random placement (which provides automation). This segmentation allows both structured placement requirements and automation to coexist by applying them at different stages of the design flow.
Solution Approach 2:
The patent applies preliminary structured placement actions before standard cell random placement to establish the regularity and organization needed in the design. By performing structured placement first, the design achieves the required regularity, and then automation is introduced in subsequent phases without compromising the established structure.
2Manufacturing precision
If manual cell sizing and optimization is performed, then placement quality is improved, but engineering effort and cost increase significantly
Solution Approach 1:
The patent implements self-service by enabling the placement tool to automatically perform cell sizing and optimization using the structured placement information and regularity patterns established earlier in the process. The system uses its own internal algorithms and the structured data it created to automatically achieve high placement quality without requiring external manual intervention, thereby reducing engineering effort while maintaining precision.
3Manufacturing precision
If ideal orientation of all cells is sought, then circuit design metrics are improved, but computational complexity becomes intractable
Solution Approach 1:
The patent applies local quality by optimizing cell orientations locally within structured placement blocks rather than globally across the entire design. By focusing optimization on local regions where structured placement rules apply, the system achieves improved circuit design metrics while avoiding the combinatorial explosion that would result from optimizing all cell orientations globally, thus keeping computational complexity manageable.
Solution Approach 2:
The patent applies partial action by optimizing orientations for only those cells that are part of structured placement blocks, rather than attempting to optimize all cells in the design. This partial optimization approach achieves sufficient improvement in circuit design metrics for the critical structured portions of the design without incurring the prohibitive computational cost of exhaustive global optimization.
4Manufacturing precision
If structured placement is maintained through multiple iterations, then placement quality is improved, but convergence difficulty increases
Solution Approach 1:
The patent performs preliminary structured placement to establish regularity and organization before initiating iterative optimization processes. By having the structured placement framework in place beforehand, subsequent iterations can focus on refining specific aspects of the design rather than rebuilding the entire structure, thereby improving placement quality while reducing the time required for convergence.
Solution Approach 2:
The patent segments the iterative optimization process into distinct phases that respect the structured placement boundaries. Each iteration focuses on optimizing specific aspects within the structured framework rather than performing exhaustive global searches, which improves placement quality while avoiding the exponential time loss that would result from unsegmented exhaustive optimization.
Data Source
AI summary
Methods and apparatuses are disclosed for automatic orientation optimization in the course of generating a placed, routed, and optimized circuit design. Also disclosed are a circuit design and circuit created with the technology. Also disclosed are a circuit design and circuit created with the technology.