Files in this item



application/pdf9503329.pdf (4MB)Restricted to U of Illinois
(no description provided)PDF


Title:Probabilistic simulation for reliability and average power estimation
Author(s):Stamoulis, Georgios Ioannis
Doctoral Committee Chair(s):Hajj, Ibrahim N.
Department / Program:Electrical and Computer Engineering
Discipline:Electrical and Computer Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Engineering, Electronics and Electrical
Abstract:The trend in the integrated circuit industry towards an ever greater miniaturization of circuit components accentuated reliability problems such as electromigration in the power buses, hot-carrier degradation of transistors, and excessive power consumption, which ultimately resulted in chip failures. This brought forward the need for simulation tools that would evaluate the susceptibility to the aforementioned hazards before the actual fabrication process.
As an answer to the above concerns, a probabilistic simulation strategy is proposed in this thesis, which provides a quick yet reliable estimate of the electromigration and hot-carrier degradation of a circuits and its average power consumption. This work expands on previously proposed concepts and introduces new ones, to establish the validity and the accuracy of the probabilistic simulation approach.
The focus of this research was on issues concerning the simulation of one channel-connected subcircuit (gate) as well as on problems related to the simulation of an entire circuit. More specifically, in the subcircuit level, a new algorithm for probabilistic graph elimination is proposed, a new delay estimation methodology, and new models for estimating the average current drawn by each gate. At the circuit level, event merging is introduced, and the problem of correlation is addressed through the study of many different approaches, all of which are introduced in this work. The scope of probabilistic simulation is expanded to sequential circuits by the introduction of an algorithm for this purpose. Finally, an extended probabilistic waveform model is introduced, a model allowing finite voltage signal slopes, which further enhances the accuracy of probabilistic analysis and lays the foundation for future work.
Issue Date:1994
Rights Information:Copyright 1994 Stamoulis, Georgios Ioannis
Date Available in IDEALS:2011-05-07
Identifier in Online Catalog:AAI9503329
OCLC Identifier:(UMI)AAI9503329

This item appears in the following Collection(s)

Item Statistics