Hierarchical Chip Design IP Identification via Cell Name Propagation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in identifying and reporting functional changes in hierarchical chip design intellectual property, as non-functional name changes introduce noise into the analysis process, making it difficult to understand the scope and impact of design changes in large-scale production designs.
Innovation Solution
The method involves renaming functionally identical cells to have matching names, which are propagated upward through the design hierarchy, using digests to detect similarities and differences, and applying canonical cell naming to reduce noise caused by non-functional name changes, allowing for the classification and filtering of such changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If cell names are changed at all levels of the design hierarchy, then design flexibility and adaptability are improved, but noise in the analysis process increases and measurement precision deteriorates
Solution Approach 1:
The patent introduces digests as an intermediary mechanism that decouples cell name changes from functional change detection. By computing digests from cell geometry and connectivity data rather than relying on cell names, the system can identify functional changes without being affected by non-functional naming variations at any hierarchy level.
Solution Approach 2:
The patent performs preliminary computation of digests for all cells before change detection. This preliminary action creates a baseline functional representation that allows subsequent comparison to distinguish functional changes from non-functional name changes, improving measurement precision without restricting design flexibility.
2Reliability
If all cells are analyzed for changes in large production scale designs, then completeness of analysis is improved, but computational complexity and time consumption increase
Solution Approach 1:
The patent segments the large design into a hierarchical structure and processes cells level by level, computing digests for leaf cells first, then propagating upward through intermediate levels to the top level. This segmentation allows the analysis to be performed on manageable subsets of cells at each hierarchy level rather than all cells simultaneously, reducing computational complexity while maintaining completeness.
Solution Approach 2:
The patent implements selective change detection by comparing digests only for cells that have changed names or have new instances, rather than re-analyzing all cells. This partial action approach maintains analysis completeness for changed elements while significantly reducing computational overhead in large production designs.
3Measurement precision
If digests are computed for all cells including higher level hierarchical cells, then functional equivalence detection is improved, but computational time and resource usage increase
Solution Approach 1:
The patent segments digest computation into hierarchical levels, computing digests for leaf cells first, then using those results to compute digests for intermediate level cells, and finally for top level cells. This bottom-up segmentation allows functional equivalence detection at each level without requiring complete re-computation of all higher levels, reducing computational time while maintaining precision.
Solution Approach 2:
The patent performs preliminary digest computation for leaf cells and stores these results before processing intermediate and top level cells. This preliminary action enables higher level digest computation to reuse lower level results rather than re-computing from scratch, improving measurement precision for functional equivalence while reducing overall computational time.
Data Source
AI summary
One method implementation disclosed includes detecting matching leaf cells that have functionally identical designs (optionally, similar designs) and assigning matching names for the matching leaf cells to replace original, non-matching names. Optionally, digests can be calculated for the leaf cells and used to detect similarities and/or differences. The matching names are propagated to at least some higher-level cells in the hierarchical design, in place of the original names. The method can further include calculating digests for at least some of the higher level cells after the propagating of the matching names into the higher level cells. Various design matching technologies can be used in combination with cell renaming and new name propagation, not limited to use of digests. Dependency chains can be calculated to improve propagation of names through the hierarchy.