Statistical Graph Circuit Component Probability Model for VLSI Design
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current electronic design automation (EDA) systems face challenges in efficiently analyzing and predicting the performance of integrated circuit designs due to the complexity of very large-scale integration (VLSI) circuit graphs, which require heavy computation and are not efficiently handled by common machine learning techniques, especially when dealing with hierarchical structures and directional edges.
Innovation Solution
The proposed method uses a statistical graph component probability model that converts local circuit graph data into tabular features, applies machine learning techniques such as regression and neural networks, and employs a neighborhood hash process to capture graph topology information, reducing computational complexity and processor resources by using a bookkeeping graph and hashing table for efficient prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If common machine learning techniques are used to analyze VLSI circuit graphs, then prediction capability is provided, but computational complexity and processor resources increase significantly
Solution Approach 1:
The patent segments the circuit graph analysis by converting local graph data into tabular features for individual components, then aggregating results. The neighborhood hash process divides the graph into manageable neighborhoods, and the bookkeeping graph organizes statistical information in a structured manner, reducing overall computational complexity while maintaining prediction accuracy.
Solution Approach 2:
The patent introduces a bookkeeping graph as an intermediary data structure that stores statistical information about circuit components and their neighborhoods. This intermediary structure enables efficient querying and prediction without repeatedly processing the entire complex circuit graph, thereby reducing computational resources required for analysis.
2Measurement precision
If detailed feature engineering is performed to improve prediction accuracy, then measurement precision improves, but processor time and complexity increase
Solution Approach 1:
The patent performs preliminary action by pre-computing and storing statistical information about circuit components and their neighborhoods in the bookkeeping graph during a training phase. This preliminary processing eliminates the need for extensive feature engineering during actual prediction, reducing processor time while maintaining high measurement precision through the pre-compiled statistical models.
3Measurement precision
If hierarchical structures and directional edges are fully considered in the analysis, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent applies local quality by analyzing specific neighborhoods around circuit components rather than the entire graph. The neighborhood hash process captures local structural characteristics including hierarchical relationships and directional edges within each neighborhood, preserving measurement precision while reducing overall device complexity by focusing computation on local rather than global properties.
Data Source
AI summary
A system and method predicts performance of a circuit design by receiving circuit design training data and circuit design test data. The circuit design training data includes training nodes and training paths. The training paths connect the training nodes including circuit components. The circuit design test data includes a first test node and a second test node. Further, testing information is determined for the circuit components of each training path from the circuit design training data. A statistical representation of the circuit design test data is determined based on the testing information and the circuit design test data, and first test information for a test path connecting the first test node with the second test node is determined based on the statistical representation.