Continuous Test Grading Flow for IC Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for verifying complex integrated circuit chips require large sets of tests, leading to inefficient resource usage due to exponential scaling of computation, memory, and time, making it costly and time-consuming to qualify changes.
Innovation Solution
A system for continuous test coverage grading that divides tests into groups, grades them independently, and analyzes results as they complete, reducing resource requirements and allowing for partial results and progress tracking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all tests in a regression run are executed and graded together using traditional methods, then complete test coverage analysis is achieved, but computation resources, memory, and time requirements scale exponentially with the number of tests
Solution Approach 1:
The patent divides the regression test run into multiple batches, processing and grading tests in smaller groups rather than all at once. Each batch is graded independently, and results are accumulated to achieve complete coverage analysis. This segmentation reduces the exponential resource scaling by breaking down the large-scale problem into manageable smaller problems that can be handled with linear resource growth.
2Measurement precision
If all tests are graded together using traditional methods, then comprehensive coverage metrics are obtained, but the process requires waiting for the entire regression run to complete before any results are available
Solution Approach 1:
The patent performs grading actions on test batches as they complete execution, rather than waiting for all tests to finish. By grading batches preliminarily and accumulating results incrementally, the system provides intermediate coverage metrics during the regression run, enabling earlier decision-making while maintaining accurate final coverage analysis.
3Measurement precision
If traditional grading methods are used, then complete test coverage analysis is performed, but disk space and memory requirements are dramatically increased
Solution Approach 1:
The patent extracts and processes test grading in smaller batches rather than retaining all test data simultaneously. By taking out subsets of tests for grading and discarding processed batch data after results are accumulated, the system achieves complete coverage analysis with significantly reduced disk space and memory requirements, avoiding the need to store all test data throughout the entire process.
4Measurement precision
If the entire regression run is executed to completion, then all coverage data is collected, but the process is extremely expensive and time consuming
Solution Approach 1:
The patent implements continuous grading where useful analysis action continues throughout the regression run rather than being deferred until completion. Test batches are graded continuously as they complete, with results accumulated in real-time. This continuous useful action maintains coverage data completeness while dramatically improving productivity by providing intermediate results and enabling earlier decision-making, reducing the effective cost and time of the verification process.
Data Source
AI summary
A system for performing efficient continuous grading flow for test coverage analysis. The system provides for continuous test coverage grading. The continuous grading flow analyzes individual tests upon completion without requiring an entire set of tests to finish. As a result, the grading process at the end of the regression run is no longer necessary, disk space and memory requirements are dramatically reduced, and partial results are produced as the grading of individual tests complete, allowing engineers to track progress and make real-time decisions based on the intermediate results. The resource requirements of our continuous flow scale linearly with the number of tests rather than exponentially as with traditional approaches.