EDA Physical Verification Clustering for Runtime Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current electronic design automation (EDA) tools do not efficiently scale physical verification runtime, particularly for antenna verification, leading to lengthy processing times due to redundant operations and limited gains from using multiple computers or processors.
Innovation Solution
The implementation of a verification runtime reduction method that clusters rules and operations for parallel processing, allowing for the elimination of redundant tasks and optimizing the assignment of clusters to processors, thereby speeding up the physical verification process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional EDA tools perform physical verification sequentially, then verification completeness is maintained, but computational runtime becomes excessively long
Solution Approach 1:
The patent segments the physical verification process by dividing design rules into distinct categories (e.g., antenna rules, DRC rules, LVS rules) and further subdividing them into operational clusters. This segmentation enables independent parallel execution of different verification operations, reducing total runtime while maintaining complete verification coverage through systematic rule coverage.
Solution Approach 2:
The patent introduces a clustering dimension to the verification process, organizing rules based on operational similarity and overlap characteristics. By adding this organizational dimension, the system can efficiently assign clusters to multiple processors, transforming a sequential one-dimensional process into a parallel multi-dimensional execution model that significantly reduces runtime.
2Productivity
If multiple computers or processors are used for verification, then parallel processing capability increases, but redundant operations occur and scaling efficiency is limited
Solution Approach 1:
The patent merges similar verification operations into clusters based on operational overlap and rule similarity. By combining redundant operations into unified clusters, the system eliminates duplicate work when assigning to multiple processors, improving parallel processing efficiency and reducing the complexity of processor assignment through a systematic clustering approach.
3Measurement precision
If antenna verification is performed with high precision, then verification accuracy improves, but processing time increases significantly
Solution Approach 1:
The patent performs preliminary clustering and categorization of antenna verification rules before execution, organizing operations by overlap characteristics and computational intensity. This preliminary organization enables the system to strategically assign different antenna check operations to appropriate processors, maintaining high verification accuracy while reducing overall runtime through optimized parallel execution.
Data Source
AI summary
A method of performing a design rule check includes clustering at least one of a plurality of rules with overlapping operations from a plurality of operations or the plurality of operations with overlapping rules from the plurality of rules. The method further includes at least one of transforming at least one of the clustered plurality of operations into a first operation group or a second operation group, or transforming at least one of the clustered plurality of rules into a first rule group or a second rule group. The method even further includes at least one of assigning at least one of the first operation group to a first processor or the second operation group to a second processor, or assigning at least one of the first rule group to the first processor or the second rule group to the second processor.