Circuit Routing Demand Smoothing for Congestion Map Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional global routing techniques in electronic design automation (EDA) often result in uneven congestion maps, leading to miscorrelations between congestion maps and detailed routing, which negatively impacts power, timing, and area of the circuit design, and require increased runtime to address bumpiness.
Innovation Solution
Implementing routing demand smoothing and routing demand spreading methods to adjust the routing demand of grid cells based on average neighboring demand and transferring overflow demand to non-blocked neighboring cells, optimizing congestion maps for detailed routing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional global routing techniques are used to route nets by dividing the circuit design into a grid of cells, then the routing process can be sped up by analyzing congestion based on grid cell congestion, but uneven congestion maps are generated leading to miscorrelations between congestion maps and detailed routing, negatively impacting power, timing, and area performance
Solution Approach 1:
The patent modifies the routing demand parameter of grid cells by smoothing operations that adjust demand values based on neighboring cells. This changes the parameter distribution to eliminate bumps and create a more uniform congestion map, resolving the contradiction between fast routing analysis and accurate congestion representation.
Solution Approach 2:
The patent performs routing demand smoothing as a preliminary action before detailed routing. By pre-adjusting the routing demand values to remove congestion map inaccuracies before the detailed routing phase, it prevents the miscorrelations that would otherwise negatively impact power, timing, and area performance.
2Manufacturing precision
If routing demand smoothing and spreading methods are implemented to adjust routing demand of grid cells, then the correlation between congestion maps and detailed routing is improved, but additional processing steps are required to calculate average neighboring demand and transfer overflow demand
Solution Approach 1:
The patent segments the routing demand adjustment process into distinct operations: calculating average neighboring demand for smoothing, and transferring overflow demand for spreading. This segmentation allows each operation to be performed independently and systematically, managing the complexity through structured decomposition of the overall process.
Solution Approach 2:
The patent uses the concept of average neighboring demand as an intermediary value to smooth routing demand. By introducing this intermediate calculation step, it creates a bridge between individual cell demands and the overall congestion picture, improving correlation without requiring complex direct adjustments.
3Reliability
If routing demand is adjusted by smoothing and spreading methods, then power, timing, and area performance is enhanced, but routing runtime may be affected by the additional calculations required
Solution Approach 1:
The patent applies routing demand smoothing and spreading as partial actions rather than complete re-routings. By making targeted adjustments to routing demand values in specific areas with congestion issues, it achieves performance improvements without the computational overhead of complete routing recalculations, thus limiting runtime impact.
Data Source
AI summary
Various embodiments provide for routing a circuit design based on adjusting a routing demand. More specifically, some embodiments implement routing demand smoothing of a grid cell, routing overflow spreading of a grid cell, or some combination of both prior to detailed routing of a circuit design, which can result in improved detailed routing over conventional routing techniques.