Lookup NU author(s): Mohamed Abufalgha,
Dr Alex Bystrov
This is the authors' accepted manuscript of a conference proceedings (inc. abstract) that has been published in its final definitive form by IEEE, 2017.
For re-use rights please refer to the publisher's terms and conditions.
© 2017 IEEE. A new method of evaluating the reliability of combinational circuits is proposed, this method uses two levels of characterisation: a Stochastic Fault Model (SFM) of the component library and a design-specific Critical Vector Model (CVM). The idea is to move the high-complexity problem of stochastic characterisation of parameters into the generic part of the design process, and do it just once for a great number of the specific designs. The SFM captures variations of the vector of parameters of a library component fault model, those causing a transient fault at the component output; it is meant to be supplied by the foundry, similar to timing library files. The CVM is derived by a limited number of simulation runs on the specific design, and represents the boundary between the erroneous and error-free operation, w.r.t. the vector of parameters of each component. The probability of error-free operation is subsequently calculated by jointly using SFM and CVM. The method is demonstrated on a chain of inverters for simplicity, and subsequently applied to a 3-bit full adder. A complex three-way trade-off between energy, performance and reliability is explored. The method is meant to serve as a basis for design-time reliability evaluation and runtime power-reliability management. A slow stage is added to the circuit under test to improve its reliability.
Author(s): Abufalgha MA, Bystrov A
Publication type: Conference Proceedings (inc. Abstract)
Publication status: Published
Conference Name: 23rd International Symposium on On-Line Testing and Robust System Design (IOLTS 2017)
Year of Conference: 2017
Online publication date: 21/09/2017
Acceptance date: 02/04/2016
Date deposited: 19/12/2017
Library holdings: Search Newcastle University Library for this item