Close category search window
 

Improving scratch-pad memory reliability through compiler-guided data block duplication

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

4 Author(s)
Li, F. ; Dept. of Comput. Sci. & Eng., Pennsylvania State Univ., University Park, PA, USA ; Chen, Gu. ; Kandemir, M. ; Kolcu, I.

Recent trends in embedded computing indicates an increasing use of scratch-pad memories (SPMs) as on-chip store for instructions and data. An important characteristic of these memory components is that they are managed by software, instead of hardware. Ever-scaling process technology and employment of several power-saving techniques in embedded systems (e.g., voltage scaling) make these systems particularly vulnerable to soft errors and other transient errors. Therefore, it is very important in practice to consider the impact of soft errors in SPMs. While it is possible to employ classical memory protection mechanisms such as parity checks and ECC, each of these has its drawbacks. Specifically, a pure parity-based protection cannot correct any errors, and ECCs can be an overkill in the normal operation state when no soft error is experienced. This paper proposes an alternate approach to protect SPMs against soft errors. The proposed approach is based on data block duplication under compiler control. More specifically, an optimizing compiler duplicates data blocks within the SPM and protects each data block by parity if such a duplication does not hurt performance. The goal of this scheme is to provide only parity protection for data blocks (and reduce the overheads at runtime when no error occurs) but correct errors using the duplicate (when an error occurs in the primary copy), provided that the duplicate is not corrupted.

Published in:
Computer-Aided Design, 2005. ICCAD-2005. IEEE/ACM International Conference on

Date of Conference: 6-10 Nov. 2005

Need Help?


IEEE Advancing Technology for Humanity About IEEE Xplore | Contact | Help | Terms of Use | Nondiscrimination Policy | Site Map | Privacy & Opting Out of Cookies

A not-for-profit organization, IEEE is the world's largest professional association for the advancement of technology.
© Copyright 2013 IEEE - All rights reserved. Use of this web site signifies your agreement to the terms and conditions.