Please use this identifier to cite or link to this item: http://scholarbank.nus.edu.sg/handle/10635/16329
Title: Memory Optimizations for Time-Predictable Embedded Software
Authors: VIVY SUHENDRA
Keywords: real-time, embedded systems, scratchpad memory, cache, WCET, performance optimization
Issue Date: 12-Aug-2009
Source: VIVY SUHENDRA (2009-08-12). Memory Optimizations for Time-Predictable Embedded Software. ScholarBank@NUS Repository.
Abstract: Real-time systems are required to accomplish their assigned functionality in a certain timeframe. This calls for worst-case performance enhancement in order to build the system with efficient resource utilization. The enhancement should not compromise timing predictability: the property that system performance can be bounded to give safe timing guarantees in all circumstances. An important optimization for embedded systems is the inclusion of fast memory on-chip. However, popular memory schemes for average-case optimization, typically profile-based, cannot effectively improve worst-case performance. This thesis proposes memory optimization techniques to improve worst-case performance while preserving time-predictability of real-time embedded software. We evaluate time-predictable shared-cache schemes to extract design guidelines given application characteristics. For scratchpad memory, an inherently time-predictable construct, we present optimal and heuristic-based allocation techniques. We handle concurrent multitasking environment by flexible space-sharing and time-multiplexing schemes based on task interactions. The thesis concludes with simultaneous global optimization of scratchpad allocation and multiprocessor scheduling.
URI: http://scholarbank.nus.edu.sg/handle/10635/16329
Appears in Collections:Ph.D Theses (Open)

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
VivySuhendra_PhDThesis_2009.pdf1.69 MBAdobe PDF

OPEN

NoneView/Download

Page view(s)

259
checked on Dec 18, 2017

Download(s)

156
checked on Dec 18, 2017

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.