Unit delay prediction method and element delay sensitivity calculation method based on neural network

Abstract

The invention discloses a unit delay prediction method based on a neural network, comprising the following steps: S1, selecting characteristic quantities required by SPICE simulation and neural network training; 2, randomly selecting that characteristic value, adopting SPICE to simulate and measure the delay of the unit, and establishing a sample set of the delay of the unit; S3, dividing the unitdelay sample set obtained by SPICE simulation into training sample set and test sample set, training the neural network model by a training sample set, and verifying the accuracy of neural network bytesting sample set, comparing the error between the unit delay predicted by testing sample set and the unit delay measured by SPICE simulation, and optimizing repeatedly the neural network parametersto reduce the error. The neural network model is a prediction model of unit delay. The invention also discloses a cell delay sensitivity calculation method. The invention has the advantages of high precision, low modeling cost and fast prediction speed.

Description

technical field [0001] The invention relates to the field of integrated circuit design, in particular to a unit delay prediction method and a unit delay sensitivity calculation method. Background technique [0002] In recent years, application-driven has gradually become a new development model for the integrated circuit industry. Due to the technical requirements of terminal products in emerging applications such as the Internet of Things, integrated circuits are developing towards smaller transistor sizes and lower operating voltages. In this year's process, the voltage has been able to be reduced to the near-threshold range, where circuit performance is largely affected by process parameter fluctuations. The analysis of unit delays with non-Gaussian distribution is a new challenge faced by the academic community. There are various limitations in the existing timing analysis methods: the static timing analysis method based on the process angle has an overly pessimistic es...

AI Technical Summary

Problems solved by technology

There are various limitations in the existing timing analysis methods: the static timing analysis method based on the process angle has an overly pessimistic estimate of the circuit delay; although the analysis method based on the Monte Carlo simulation can accurately simulate the circuit delay, the simulation run The quantity is large and the operation takes a long time, so it cannot be used as a method for circuit timing analysis

Images