Sequential redundancy identification using transformation-based verification


Sequential redundancy identification using transformation-based verification

Show full record

Title: Sequential redundancy identification using transformation-based verification
Author: Mony, Hari, 1977-
Abstract: The design of complex digital hardware is challenging and error-prone. With short design cycles and increasing complexity of designs, functional verification has become the most expensive and time-consuming aspect of the digital design process. Sequential equivalence checking (SEC) has been proposed as a verification framework to perform a true sequential check of input/output equivalence between two designs. SEC provides several benefits that can enable a faster and more efficient way to design and verify large and complex digital hardware. It can be used to prove that micro-architectural optimizations needed for design closure preserve design functionality, and thus avoid the costly and incomplete functional verification regression traditionally used for such purposes. Moreover, SEC can be used to validate sequential synthesis transformations and thereby enable design and verification at a higher-level of abstraction. Use of sequential synthesis leads to shorter design cycles and can result in a significant improvement in design quality. In this dissertation, we study the problem of sequential redundancy identification to enable robust sequential equivalence checking solutions. In particular, we focus on the use of a transformation-based verification framework to synergistically leverage various transformations to simplify and decompose large problems which arise during sequential redundancy identification to enable an efficient and highly scalable SEC solution. We make five main contributions in this dissertation. First, we introduce a novel sequential redundancy identification framework that dramatically increases the scalability of SEC. Second, we propose the use of a flexible and synergistic set of transformation and verification algorithms for sequential redundancy identification. This more general approach enables greater speed and scalability and identifies a significantly greater degree of redundancy than previous approaches. Third, we introduce the theory and practice of transformation-based verification in the presence of constraints. Constraints are pervasively used in verification testbenches to specify environmental assumptions to prevent illegal input scenarios. Fourth, we develop the theoretical framework with corresponding efficient implementation for optimal sequential redundancy identification in the presence of constraints. Fifth, we address the scalability of transformation-based verification by proposing two new structural abstraction techniques. We also study the synergies between various transformation algorithms and propose new strategies for using these transformations to enable scalable sequential redundancy identification.
Department: Electrical and Computer Engineering
Subject: Computer software--Verification Digital electronics
Date: 2008-05

Files in this work

Download File: monyh96808.pdf
Size: 679.1Kb
Format: application/pdf

This work appears in the following Collection(s)

Show full record

Advanced Search


My Account