Circuit Model Hashing for Verification Redundancy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Model checking of complex circuit designs is time-consuming and inefficient, requiring repeated verification of circuit design changes due to the exhaustive nature of formal verification methods.
Innovation Solution
A method utilizing a centralized persistent database to cache results of prior model checking runs, where circuit models are hashed and compared to identify equivalent designs, allowing for the retrieval of prior test results and avoiding redundant model checks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If formal verification (model checking) is performed exhaustively on circuit designs, then verification reliability is improved, but verification time increases significantly
Solution Approach 1:
The system performs preliminary hashing of circuit models before exhaustive model checking. By computing hash values of circuit netlists and comparing them against a database of previously checked circuits, the system can identify equivalent circuits in advance and retrieve their verification results, avoiding redundant exhaustive model checking while maintaining verification reliability.
Solution Approach 2:
The system creates hash representations (copies) of circuit models that serve as fingerprints for identification. These hash copies are stored in a database and used to quickly identify equivalent circuits without needing to perform full model checking again, thus reducing verification time while preserving reliability through the use of actual verification results from equivalent circuits.
2Manufacturing precision
If iterative circuit design changes are made and each iteration is model checked, then design accuracy is improved, but design productivity deteriorates
Solution Approach 1:
The system performs preliminary hash comparison checks before initiating full model checking iterations. When designers make iterative changes to circuit designs, the system first hashes the modified netlist and checks the database for equivalent circuits. If an equivalent is found, the existing verification results are reused, allowing designers to iterate quickly while maintaining accuracy through selective re-verification of only truly changed designs.
Solution Approach 2:
The system changes the parameter representation of circuit models by converting detailed netlists into compact hash values. This parameter transformation enables rapid comparison and identification of equivalent circuits across design iterations, allowing the system to distinguish between meaningful changes and redundant verifications, thereby improving design productivity without sacrificing accuracy.
3Reliability
If exhaustive model checking is performed on complex circuit designs, then verification completeness is improved, but computational resources are excessively consumed
Solution Approach 1:
The system extracts the essential identifying特征 of circuit models by computing hash values from netlists, separating the identification function from the exhaustive verification function. This extraction allows the system to quickly identify equivalent circuits and retrieve their verification results without re-consuming computational resources on already-verified designs, thus reducing resource consumption while maintaining verification completeness.
Solution Approach 2:
The system performs preliminary hash-based identification and database lookup before initiating resource-intensive exhaustive model checking. This preliminary action filters out redundant verification cases by identifying equivalent circuits in advance, ensuring that computational resources are consumed only when necessary for truly novel or modified circuit designs, thereby reducing overall resource consumption while preserving verification completeness.
Data Source
AI summary
A method, system or computer usable program product for model checking a first circuit model including receiving a request from a user for a model check of the first circuit model; responsive to receiving the user request, simulating the first circuit model to generate simulation results; hashing the first circuit model simulation results to generate a hash index; comparing the hash index to a database of prior hash indices generated from hashed simulation results of prior circuit models to determine whether the first circuit model hash index matches a prior hash index of any of the prior circuit models to identify a matching prior circuit model; upon a positive match, determining whether the first circuit model is equivalent to the matching prior circuit model; and upon a positive determination of equivalence, providing prior test results of the matching prior circuit model to the user.