Integrated circuit simulation data correlation modeling method and device

Abstract

The invention belongs to the field of integrated circuit design automation, and specifically relates to a method and device for modeling the correlation of integrated circuit simulation data based on correlation clustering and covariance shrinkage technology: in the invention, the circuit simulation data required for modeling is firstly obtained, Then an original multivariate normal distribution is constructed according to the data, and the distribution is corrected by correlation clustering and covariance shrinkage techniques to obtain a correlation model represented by the revised multivariate normal distribution. The invention improves the reliability and accuracy of the correlation model, so that the correlation model of circuit simulation data can be applied to circuits of any scale, and the accuracy and efficiency of the algorithm developed by using the model can be guaranteed.

Description

technical field [0001] The invention belongs to the field of integrated circuit design automation, and in particular relates to modeling of correlation of integrated circuit simulation data. Specifically, it relates to a correlation modeling method and device for integrated circuit simulation data based on correlation clustering and covariance shrinkage techniques. Background technique [0002] The prior art discloses that since integrated circuits are inevitably affected by process conditions during the manufacturing process, the physical parameters of each transistor, such as doping concentration, gate oxide layer thickness, surface charge, etc., will fluctuate randomly. Therefore, the electrical properties of the transistor Characteristics such as threshold voltage, leakage current, etc. and circuits composed of these transistors also have a certain degree of randomness in performance (such as delay, gain, etc.). With the development of integrated circuit process technol...

AI Technical Summary

Problems solved by technology

For a given set of multivariate data, the correlation of any two dimensions can be determined simply by unbiased estimation or maximum likelihood estimation of the covariance matrix of the multivariate data; however, due to the limitation of simulation cost, the simulation of integrated circuits The data often have the characteristics of high dimensionality and few samples. In this case, the estimated value of the covariance matrix is ​​unstable and has a large error, which is usually not directly used for computer processing.

Images