|
Abstract: |
Voltage scaling has been extensively used in industry for decades to reduce power consumption. In recent years, exploring digital circuit operation in moderate inversion has created an interest among researchers due to its immense capability to provide a perfect tradeoff between high performance and low energy operation. But circuits operating in moderate inversion are susceptible to process variations and variability. To compute variability, statistical parameters such as the probability density function (PDF) and cumulative distribution function (CDF) are required. This paper presents an analytical model framework for delay calculations utilizing log skew normal distribution for ultradeep submicron technology nodes up to 22 nm. The CDF of the proposed model is utilized to calculate minimum and maximum delays with 3σ-accuracy providing better accuracy than the conventional methods. The obtained results are also compared with Monte Carlo simulations with errors lying within the acceptable range of 2%-4%. |
Key words: moderate inversion ultradeep submicron predictive technology model variability log skew normal distribution |
DOI:10.11916/j.issn.1055-9113.2022023 |
Clc Number:TN91,TK01 |
Fund: |