Gradient AOCV Tables for IC Timing Closure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Graph-Based Analysis (GBA) methods in integrated circuit design are overly pessimistic, leading to increased area and power requirements due to inaccurate timing analysis, making it difficult to achieve timing closure and resulting in inefficient use of resources.
Innovation Solution
The creation of custom, k-Level Gradient Advanced On-Chip Variation (AOCV) tables based on design depth profiles, which reduces pessimism by setting derate values for shorter paths to match longer paths, enabling more accurate delay estimation and improving correlation between Path-Based Analysis (PBA) and GBA timing numbers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional GBA methods use depth-dependent derates for all paths, then timing analysis coverage is improved, but pessimism increases leading to larger area and power requirements
Solution Approach 1:
The patent applies local quality by differentiating derate application based on path characteristics. Critical paths receive depth-dependent derates for accurate timing coverage, while non-critical paths use reduced or zero derates. This localized differentiation eliminates unnecessary pessimism in non-critical paths, reducing area and power requirements while maintaining timing analysis coverage for critical paths.
Solution Approach 2:
The patent changes the derate parameter dynamically based on path criticality and depth characteristics. Instead of applying uniform depth-dependent derates to all paths, the system adjusts derate values selectively - applying full derates only where timing closure is genuinely at risk, and reducing or eliminating derates elsewhere. This parameter optimization resolves the contradiction between coverage and resource efficiency.
2Reliability
If traditional GBA methods use depth-dependent derates for all paths, then timing analysis coverage is improved, but power consumption increases
Solution Approach 1:
The patent applies local quality by differentiating derate application based on path characteristics. Critical paths receive depth-dependent derates for accurate timing coverage, while non-critical paths use reduced or zero derates. This localized differentiation eliminates unnecessary pessimism in non-critical paths, reducing area and power requirements while maintaining timing analysis coverage for critical paths.
Solution Approach 2:
The patent changes the derate parameter dynamically based on path criticality and depth characteristics. Instead of applying uniform depth-dependent derates to all paths, the system adjusts derate values selectively - applying full derates only where timing closure is genuinely at risk, and reducing or eliminating derates elsewhere. This parameter optimization resolves the contradiction between coverage and resource efficiency.
3Device complexity
If GBA uses uniform derates without depth consideration, then computational complexity is reduced, but timing analysis accuracy deteriorates
Solution Approach 1:
The patent segments the timing analysis process into two distinct phases: a fast GBA phase using uniform or reduced derates for initial timing closure, and a targeted PBA phase using depth-dependent derates only for critical paths. This segmentation maintains low computational complexity in the main flow while achieving high accuracy where needed, resolving the contradiction between complexity and precision.
Solution Approach 2:
The patent applies partial action by using simplified derate models for the majority of non-critical paths, and applying full depth-dependent derates only partially to identified critical paths. This selective application of complexity achieves sufficient timing analysis accuracy without incurring the full computational cost of applying complex models universally.
4Measurement precision
If GBA and PBA use different derate models, then timing closure accuracy is improved, but correlation between methods deteriorates
Solution Approach 1:
The patent segments the timing analysis process into two distinct phases: a fast GBA phase using uniform or reduced derates for initial timing closure, and a targeted PBA phase using depth-dependent derates only for critical paths. This segmentation maintains low computational complexity in the main flow while achieving high accuracy where needed, resolving the contradiction between complexity and precision.
Solution Approach 2:
The patent applies partial action by using simplified derate models for the majority of non-critical paths, and applying full depth-dependent derates only partially to identified critical paths. This selective application of complexity achieves sufficient timing analysis accuracy without incurring the full computational cost of applying complex models universally.
Data Source
AI summary
A method of manufacturing semiconductor circuits seeks timing closure on a preliminarily select, placed and routed set of cells using a delay for each cell as derated by a derate value obtained from a timing model table having a derate value corresponding to a circuit path depth in the netlist. The derate value for a predetermined number of circuit path depths below k are identical. The derate values are monotonically decreasing for increasing circuit depths in a range between 1.0 and 1.5. Separate timing model tables with differing identical values can be employed for standard and clock tree cells.