Circuit Reliability Simulation via Discrete Parameter Points
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current simulation methods for electronic circuits under variability, such as those affected by bias temperature instability (BTI) and random telegraph noise, are inefficient for high-sigma evaluations due to their reliance on Monte Carlo simulations, which are computationally intensive and inaccurate for non-normal distributions, especially when assessing large circuits over extended lifetimes.
Innovation Solution
A fast processor-level simulation methodology that creates a finite set of circuit parameter points within an n-dimensional space, determining response values and probabilities of occurrence for these points, allowing for a relaxed calculation of total probability and enabling accurate evaluation of circuit performance under workload-dependent degradation without relying on Monte Carlo methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If Monte Carlo simulations are used for evaluating circuit reliability under variability, then comprehensive coverage of parameter distributions is achieved, but computational complexity and simulation time increase significantly
Solution Approach 1:
The patent segments the continuous parameter space into a finite set of discrete parameter points. Instead of sampling continuously as in Monte Carlo methods, the invention divides the n-dimensional parameter space into manageable discrete points, allowing reliability assessment without requiring billions of random samples. This segmentation enables accurate high-sigma evaluation while dramatically reducing computational time.
Solution Approach 2:
The invention transforms the simulation approach by changing from random sampling parameters to systematically selected finite parameter points. By determining response values and probabilities of occurrence for these specific points, the method achieves accurate reliability assessment without the computational burden of traditional Monte Carlo simulations, particularly for high-sigma evaluations.
2Measurement precision
If Monte Carlo simulations are used for high-sigma evaluations, then statistical accuracy is maintained, but computational requirements become prohibitively intensive
Solution Approach 1:
The patent applies partial action by evaluating only the necessary finite set of parameter points rather than performing exhaustive Monte Carlo sampling. For high-sigma evaluations, the invention determines response values for specific parameter points and calculates total probability of occurrence, achieving the required precision without the excessive computational resources that would be needed for full Monte Carlo coverage.
3Reliability
If conventional simulation methods are used for non-normal distributions, then statistical rigor is maintained, but accuracy deteriorates for high-sigma evaluations
Solution Approach 1:
The patent substitutes the mechanical Monte Carlo sampling system with a deterministic system that directly calculates response values for finite parameter points. This substitution replaces random sampling with systematic evaluation, achieving superior accuracy for high-sigma evaluations and non-normal distributions by directly computing the total probability of occurrence from the finite set of parameter points rather than relying on statistical convergence.
4Adaptability or versatility
If full parameter space is simulated, then complete coverage of all possible conditions is achieved, but device complexity and computational burden increase
Solution Approach 1:
The patent extracts only the essential finite set of parameter points from the complete parameter space that are necessary for accurate reliability assessment. By determining response values and probabilities for this selected subset rather than simulating the entire continuous parameter space, the invention achieves adequate coverage of all possible conditions while significantly reducing the complexity of the simulation model.
Data Source
AI summary
A system and method for simulating an electronic circuit is disclosed. The method includes creating a finite set of circuit or device parameter points selected from within an n-dimensional parameter space. The method includes determining, for each circuit or device parameter point of the set, a corresponding response value of the performance metric and a corresponding probability of occurrence. The method includes determining, for a predetermined value of the performance metric, the total probability of occurrence.