Empirical Bayes Cell Density Control for EDA Placement Congestion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing component placement methods in integrated circuits struggle to efficiently determine and apply cell density constraints, leading to placement congestion and degradation of circuit performance.
Innovation Solution
The proposed solution involves using empirical Bayes mechanisms to adapt target cell densities based on post-route optimized outputs from Electronic Design Automation (EDA) systems, thereby improving correlation with long-running heuristics of the EDA system's router and timing-optimizer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional incremental congestion estimation and mitigation methods are used, then placement congestion can be addressed, but computational cost and time consumption increase significantly
Solution Approach 1:
The patent applies preliminary action by learning optimal cell density targets from post-route optimized outputs before actual placement. The system pre-processes routing and timing optimization results to extract density information, then uses this learned density as a constraint during placement. This eliminates the need for incremental congestion estimation during placement iterations, significantly reducing computational time while maintaining effective congestion mitigation.
2Productivity
If cell density is increased to improve placement efficiency, then placement speed improves, but placement congestion increases and circuit performance degrades
Solution Approach 1:
The patent implements feedback by using post-route optimized outputs (which reflect actual routing and timing performance) to learn optimal cell density targets. This learned density information is fed back as a constraint to guide placement, creating a closed-loop system that automatically adjusts density to achieve both high placement efficiency and good circuit performance without manual intervention.
Solution Approach 2:
The patent changes the density parameter dynamically by learning optimal density values from post-route results and applying them as constraints during placement. Instead of using fixed or uniformly incremented density values, the system adapts the density parameter based on actual routing and timing optimization outcomes, enabling simultaneous achievement of high placement efficiency and good circuit performance.
3Reliability
If cell density constraints are applied to mitigate placement congestion, then circuit performance improves, but the complexity of determining appropriate density constraints increases
Solution Approach 1:
The patent applies self-service by enabling the placement system to automatically learn optimal cell density constraints from post-route optimized outputs without requiring external input or manual configuration. The system extracts density information from routing and timing optimization results, processes it through a learning mechanism, and automatically applies the learned density as a constraint, eliminating the need for complex manual density determination procedures.
Data Source
AI summary
Mechanisms to control cell density in circuit layouts to mitigate placement congestion that learn from post-route target outputs of an Electronic Design Automation system and implement an empirical Bayes mechanisms to adapt the target to a specific component placer's achievable outputs. The disclosed mechanisms solve for component placement on a global scale and obviate the application of incremental congestion estimation and mitigation.