Circuit Simulation Device Evaluation Using Interpolated Table Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing complexity of transistor devices due to new physical effects in process technology leads to a higher number of unique device instances, resulting in increased costs for generating precharacterized table models, which in turn raises requirements for computational and storage capabilities in device calibration and circuit simulation.
Innovation Solution
A method and system for evaluating devices during circuit simulation by using a table to store input and stored values, where linear interpolation is performed between neighboring table entries when the input value does not directly map to an entry, ensuring accuracy criteria are met, thereby reducing the need for extensive precharacterization and computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If precharacterization tables are generated for all unique device instances to ensure accurate circuit simulation, then measurement precision is improved, but device complexity and storage requirements increase
Solution Approach 1:
The patent segments the device characterization space by identifying and separating corner cases (edge-case device instances) from typical device instances. This segmentation allows the system to handle only the critical corner cases through detailed precharacterization while using simplified models for typical cases, thereby reducing overall table complexity while maintaining accuracy where it matters most.
Solution Approach 2:
The patent applies local quality by providing high-resolution precharacterization data only for corner case device instances where accuracy is critical, while using lower-resolution or interpolated data for typical device instances. This localized approach to data quality reduces the overall complexity and storage requirements while maintaining measurement precision for the most challenging cases.
2Measurement precision
If precharacterization tables are generated for all unique device instances to ensure accurate circuit simulation, then measurement precision is improved, but storage capability requirements increase
Solution Approach 1:
By segmenting device instances into corner cases and typical cases, the patent reduces storage requirements by only maintaining detailed precharacterization data for the small subset of corner cases. The majority of typical device instances can be handled through interpolation or simplified models, dramatically reducing the quantity of stored data while preserving accuracy for critical cases.
Solution Approach 2:
The patent uses corner case precharacterization data as reference copies to generate or approximate data for typical device instances through interpolation or model-based approaches. This copying strategy avoids storing redundant full-characterization data for every device instance, reducing storage capacity requirements while maintaining measurement precision.
3Measurement precision
If extensive precharacterization is performed for all device instances to handle increased device complexity, then measurement precision is improved, but computational capability requirements increase
Solution Approach 1:
The patent segments the computational workload by performing extensive precharacterization computations only for corner case device instances, while using lighter computational approaches (interpolation, simplified models) for typical device instances during circuit simulation. This segmentation significantly reduces the computational power required during simulation while maintaining measurement precision for critical cases.
Solution Approach 2:
The patent performs necessary heavy computational work in advance by precharacterizing only the corner case device instances before circuit simulation begins. This preliminary action eliminates the need for repeated heavy computations during simulation, reducing real-time computational power requirements while ensuring accuracy for the most challenging device instances.
Data Source
AI summary
A method for evaluating a device during circuit simulation includes receiving a first request including a first input value; and mapping the first input value to a first space in a table. The table is configured to store one or more table entries. A table entry includes an input value and a stored value. The stored value is obtained as a function of the input value from an analytical device model used to characterize the device during circuit simulation. The method includes determining whether the first space in the table is occupied by a first table entry; selecting a second table entry and a third table entry from the table when the first space in the table is not occupied by the first table entry, wherein the second table entry and the third table entry are neighbors in relation to the first space in the table; determining whether an accuracy criteria for performing a linear interpolation is satisfied as a function of the first input value, a stored value for the second table entry, and a stored value for the third table entry; and obtaining a first response value by performing the linear interpolation of the stored value for the second table entry and the stored value for the third table entry when the accuracy criteria is satisfied; and sending a response to the first request including the first response value. The first response value corresponds with the performance of the device when the device is configured using the first input value.