ML-Based Regression Test Automation for IC Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing complexity of systems, such as integrated circuits, leads to lengthy and costly verification processes during regression testing, requiring significant computing resources and personnel due to the large number of test cases involved.
Innovation Solution
A machine learning-based automation framework is employed to classify test cases into fail classes, generate renewed test cases, and automate the simulation process by using a controller to apply solutions to test cases, thereby reducing the time and resources needed for verification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual verification of each test case is performed, then verification accuracy is maintained, but time consumption and resource requirements increase significantly
Solution Approach 1:
The patent replaces manual verification processes with an automated machine learning-based system. The ML model automatically classifies test cases into fail classes and generates renewed test cases, eliminating the need for manual intervention while maintaining verification accuracy. This substitution of mechanical manual work with automated intelligent systems directly addresses the time consumption issue.
Solution Approach 2:
The verification system performs self-service through automated classification and test case generation. The machine learning model autonomously analyzes simulation logs, identifies fail classes, and generates renewed test cases without human intervention. This self-service capability allows the system to maintain high verification accuracy while dramatically reducing time consumption and resource requirements.
2Reliability
If all test cases are simulated to ensure comprehensive verification, then verification completeness is improved, but computing resources and personnel requirements increase
Solution Approach 1:
The patent extracts and isolates the critical task of identifying fail classes from the broader verification process. By using machine learning to automatically classify test cases into fail classes and generate renewed test cases, the system extracts only the essential verification work needed, eliminating redundant manual analysis while maintaining verification completeness.
Solution Approach 2:
The system changes the parameter of verification approach from manual simulation of all test cases to automated ML-based classification and generation. This parameter change enables the system to maintain verification completeness while significantly reducing computing resources and personnel requirements through intelligent automation.
3Ease of operation
If traditional manual verification methods are used, then process simplicity is maintained, but productivity decreases due to lengthy verification processes
Solution Approach 1:
The patent replaces manual verification operations with an automated machine learning system that handles test case classification and generation. This substitution maintains operational simplicity from the user perspective while dramatically improving productivity through automated processing speeds, eliminating the trade-off between simplicity and efficiency.
Solution Approach 2:
The machine learning model serves as an intermediary between the verification system and manual operations. It automatically processes test cases, classifies fail classes, and generates renewed test cases, acting as an intelligent mediator that maintains ease of operation while significantly enhancing verification speed and productivity.
Data Source
AI summary
A method of simulating an integrated circuit includes providing at least one test case to a simulation tool, obtaining at least one first simulation result and at least one first simulation log from the simulation tool, classifying, with a first machine learning model, the at least one test case into one fail class of a plurality of fail classes, generating at least one renewed test case by applying, with a controller, a solution to the at least one test case, and providing the at least one renewed test case to the simulation tool.