Incremental Criticality and Yield Gradient Computation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for computing criticality and yield gradient in statistical static timing analysis are computationally intensive, making them inefficient for optimizing large circuits.
Innovation Solution
The method involves modeling a circuit as a directed acyclic timing graph, computing chip slack, and incrementally updating diagnostic metrics such as criticality and yield gradient, minimizing computational intensity by only updating timing results affected by changes in the circuit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods are used to compute criticality and yield gradient, then measurement precision is improved, but productivity deteriorates due to computational intensity
Solution Approach 1:
The patent segments the circuit into multiple blocks and divides the computation of criticality and yield gradient into incremental updates. Instead of computing metrics for the entire circuit simultaneously, the method processes individual blocks and their interactions separately, allowing parallel computation and reducing overall computational burden while maintaining accuracy.
Solution Approach 2:
The patent performs preliminary computation of block-level criticality and yield gradient metrics before assembling the complete circuit analysis. By pre-computing metrics for individual blocks and storing them for later retrieval and combination, the method avoids redundant calculations when analyzing the complete circuit, significantly improving optimization speed.
2Measurement precision
If complete circuit analysis is performed, then measurement precision is improved, but loss of time increases due to computational intensity
Solution Approach 1:
The patent segments the timing analysis into block-level analyses that can be performed independently and then combined. Each block's timing characteristics are analyzed separately, and the results are aggregated to produce the complete circuit timing analysis, reducing the time required for comprehensive analysis.
Solution Approach 2:
The patent implements incremental updates that allow the timing analysis to be continuously refined as circuit modifications are made. Instead of performing complete re-analysis, the method continuously updates only the affected portions of the timing analysis, maintaining accuracy while minimizing computation time throughout the optimization process.
Data Source
AI summary
In one embodiment, the invention is a method and apparatus for incrementally computing criticality and yield gradient. One embodiment of a method for computing a diagnostic metric for a circuit includes modeling the circuit as a timing graph, determining a chip slack for the circuit, determining a slack of at least one diagnostic entity, and computing a diagnostic metric relating to the diagnostic entity(ies) from the chip slack and the slack of the diagnostic entity(ies).


