Physical Verification Violation Management via Error Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional physical verification tools struggle to distinguish between real errors and intended violations, leading to potential manufacturing problems even after the circuit passes the verification stage, due to the difficulty in separating these issues, especially when a large number of violations are generated.
Innovation Solution
A system that manages violations and error classifications during physical verification by allowing users to classify DRC violations through a graphical user interface, storing these classifications, and determining how to handle them based on user input, including options to waive, watch, or ignore violations, with authorization controls to ensure accurate handling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional physical verification tools generate a large number of violations, then the verification coverage is improved, but the ability to separate real errors from intended violations deteriorates
Solution Approach 1:
The patent segments violations into different categories (real errors vs. intended violations) using error classifications. Each violation is tagged with metadata indicating its nature, allowing the system to segment and filter violations based on their classification rather than treating them as a homogeneous set. This segmentation enables better error detection while maintaining comprehensive verification coverage.
Solution Approach 2:
The patent introduces an intermediary classification layer between the violation detection mechanism and the user. Error classifications act as intermediaries that provide context about each violation's nature, serving as a mediator that helps users distinguish between real errors and intended violations without requiring direct analysis of each violation's technical details.
2Measurement precision
If users manually review each violation to determine error classification, then the accuracy of error identification is improved, but the verification time increases substantially
Solution Approach 1:
The system performs preliminary actions by automatically generating error classifications and metadata for each violation before user review. This preliminary classification work reduces the cognitive load on users and enables them to quickly identify real errors without manually analyzing each violation from scratch, thereby maintaining accuracy while reducing verification time.
Solution Approach 2:
The violation detection system serves itself by automatically generating error classifications, metadata, and contextual information for each violation. This self-service capability reduces the manual effort required for error identification while maintaining high accuracy, as the system prepares structured information that guides user decision-making.
3Reliability
If the system applies design rule checks to all cells in the layout, then the verification completeness is improved, but the processing speed deteriorates
Solution Approach 1:
The patent applies local quality by using error classifications to differentiate verification intensity for different cells. Cells with known error patterns or intended violations receive simplified verification, while cells with unknown or critical violations receive full verification. This localized approach maintains verification completeness for critical areas while speeding up processing for routine areas.
Solution Approach 2:
The system performs preliminary verification on cells by checking against stored error classifications and patterns before applying full design rule checks. This preliminary action identifies cells that can be quickly verified or skipped, maintaining verification completeness for cells that require it while dramatically speeding up processing for cells with known characteristics.
4Productivity
If the system stores and reuses error classifications for matching cells, then the verification speed is improved, but the memory requirements increase
Solution Approach 1:
The patent uses copying by storing error classifications and metadata from verified cells and reusing them for matching cells. Instead of re-verifying identical or similar cells, the system copies previously determined error classifications and applies them to matching cells, dramatically improving verification speed while requiring minimal additional memory storage.
Data Source
AI summary
Some embodiments provide a system for managing violations during physical verification. The system may identify a design-rule-check (DRC) violation by applying a set of DRC rules to a layout. The system can then receive an error classification from the user which specifies how the DRC violation is to be handled. Next, the system can store the DRC violation, the user-selected error classification, and a user identifier associated with the user in a database. If the user is not authorized to approve the error classification, the database can indicate that the error classification has not been approved. Later, a user who is authorized to approve the error classification can approve the error classification. The system can determine if a cell is known, and if so, the system can use the violations and error classifications stored in the database to speed up the verification process.