Random Sub-Space Generation for IC Verification Coverage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current design verification methods, particularly for integrated circuits and FPGA devices, face inefficiencies in testing due to the need for exhaustive permutation sets of parameters, which are not scalable and require extensive simulation time, often covering unimportant or redundant parameter combinations.
Innovation Solution
A method and system for generating a random sub-space of assignments for generative variables based on user-defined distribution traits, allowing for focused testing on specific parameters and sub-groups, reducing the number of permutations and simulation time by prioritizing important parameter combinations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If exhaustive permutation sets of parameters are used for verification, then verification coverage is improved, but simulation time and computational resources increase significantly
Solution Approach 1:
The patent segments the complete parameter space into multiple sub-spaces, each characterized by specific distribution traits. Instead of exhaustively testing all permutations, the system divides the verification task into manageable segments that can be tested independently, reducing overall simulation time while maintaining coverage of critical parameter combinations.
Solution Approach 2:
The system changes the parameters of test cases by applying user-defined distribution traits to generate varied parameter sets. This allows the verification to explore different regions of the parameter space systematically, achieving good coverage without requiring exhaustive testing of all possible permutations.
2Reliability
If all parameter permutations are tested, then verification completeness is improved, but device complexity and test generation complexity increase
Solution Approach 1:
The system performs preliminary action by pre-defining distribution traits and characterizing sub-spaces before actual verification begins. This upfront preparation work organizes the parameter space into structured sub-spaces with specific characteristics, making the subsequent verification process simpler and more manageable rather than dealing with the full complexity of all permutations.
Solution Approach 2:
The patent applies partial action by selecting and testing only certain sub-spaces that are most relevant to verification goals, rather than exhaustively testing all parameter permutations. The system generates a sufficient number of test cases from selected sub-spaces to achieve adequate coverage without the excessive complexity of complete permutation testing.
3Productivity
If random sub-space generation with distribution traits is used, then test generation efficiency is improved, but coverage of certain parameter combinations may be reduced
Solution Approach 1:
The system incorporates feedback mechanisms where distribution traits are defined based on analysis of the device under test and verification goals. The feedback loop allows the system to learn from testing results and adjust the selection and characterization of sub-spaces, ensuring that efficient random generation does not compromise coverage of important parameter combinations.
Solution Approach 2:
The patent applies local quality by assigning different distribution traits to different sub-spaces based on their specific characteristics and importance. Rather than using a uniform random generation approach, the system tailors the distribution characteristics to each sub-space, ensuring that critical parameter combinations receive appropriate attention while maintaining overall test generation efficiency.
Data Source
AI summary
System, method and computer readable medium are described. The method may include obtaining user defined distribution traits characterizing a random sub-space of a space of assignments for a set of generative variables. The method may further include applying the user defined distribution traits on the space of assignments for a set of generative variables to generate the random sub-space of the space of assignments for a set of generative variables. The method may also include testing a device under test using the generated random sub-space of the space of assignments for a set of generative variables.