Partitioning in hardware software codesign of embedded systems for minimization of power /
Embedded systems are seen almost everywhere in day to day life starting from toys to automobiles. These systems generally consist of both hardware and software components. The traditional design approach of ES has been somewhat inflexible and error prone in terms of exploring the vast design space t...
| Main Author: | |
|---|---|
| Format: | Thesis Book |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
2001.
|
| Subjects: |
| Summary: | Embedded systems are seen almost everywhere in day to day life starting from toys to automobiles. These systems generally consist of both hardware and software components. The traditional design approach of ES has been somewhat inflexible and error prone in terms of exploring the vast design space to arrive at an optimal design. This is intolerable when the embedded systems face the need of a reduced time to market due to short life span. The growing demand on a short design cycle is met by using a new design methodology called Hardware Software Codesign. In hardware-software codesign, the designers consider the design of hardware and software blocks concurrently until prototype implementation. In the codesign framework, the step on partitioning is important. A properly partitioned design gives a better product, that satisfies cost and performance and meets other constraints. Power consumption in these ubiquitous devices is of major concern for designers and users as well. In this thesis we consider the power consumption parameter as our focus while partitioning the HW/SW codesign. We propose a modified codesign framework that emphasizes power as an important parameter in the partitioning step. Our algorithm EGLT is an exhaustive method to provide minimum power as the result. The algorithm can be changed to minimize other parameters such as area and time too. Although the partitioning problem is known to be NP complete, the EGLT based partitioning is found to be reasonably efficient for medium sized embedded system design. |
|---|---|
| Item Description: | Vita. "Major subject: Computer Engineering". |
| Physical Description: | xi, 106 leaves : illustrations ; 28 cm. Issued also on microfiche from Lange Micrographics. |
| Bibliography: | Includes bibliographical references (leaves 96-98). |