Hardware/software partitioning of streaming applications for multi-processor system-on-chip
Hardware/software (HW/SW) co-design has emerged as a crucial and integral part in the development of various embedded applications. Moreover, the increases in the number of embedded multimedia and medical applications make streaming throughput an important attribute of Multi-Processor System-on-Chip...
Saved in:
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
ARQII Publication, Malaysian Simulation Society
2017
|
Subjects: | |
Online Access: | http://eprints.utm.my/id/eprint/80739/1/MuhammadNadzirMarsono2017_HardwareSoftwarePartitioningofStreaming.pdf http://eprints.utm.my/id/eprint/80739/ http://arqiipubl.com/ojs/index.php/AMS_Journal/article/view/2/2 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
my.utm.80739 |
---|---|
record_format |
eprints |
spelling |
my.utm.807392019-06-27T06:20:54Z http://eprints.utm.my/id/eprint/80739/ Hardware/software partitioning of streaming applications for multi-processor system-on-chip Tang, J. W. Hau, Y. W. Marsono, M. N. TK Electrical engineering. Electronics Nuclear engineering Hardware/software (HW/SW) co-design has emerged as a crucial and integral part in the development of various embedded applications. Moreover, the increases in the number of embedded multimedia and medical applications make streaming throughput an important attribute of Multi-Processor System-on-Chip (MPSoC). As an important development step, HW/SW partitioning affects the system performance. This paper formulates the optimization of HW/SW partitioning aiming at maximizing streaming throughput with predefined area constraint, targeted for multi-processor system with hardware accelerator sharing capability. Software-oriented and hardware-oriented greedy heuristics for HW/SW partitioning are proposed, as well as a branch-and-bound algorithm with best-first search that utilizes greedy results as initial best solution. Several random graphs and two multimedia applications (JPEG encoder and MP3 decoder) are used for performance benchmarking against brute force ground truth. Results show that the proposed greedy algorithms produce fast solutions which achieve 87.7% and 84.2% near-optimal solution respectively compared to ground truth result. With the aid of greedy result as initial solution, the proposed branch-and-bound algorithm is able to produce ground truth solution up to 2.4741e+8 times faster in HW/SW partitioning time compared to exhaustive brute force method. ARQII Publication, Malaysian Simulation Society 2017 Article PeerReviewed application/pdf en http://eprints.utm.my/id/eprint/80739/1/MuhammadNadzirMarsono2017_HardwareSoftwarePartitioningofStreaming.pdf Tang, J. W. and Hau, Y. W. and Marsono, M. N. (2017) Hardware/software partitioning of streaming applications for multi-processor system-on-chip. Applications of Modelling and Simulation, 1 (1). pp. 1-14. ISSN 2600-8084 http://arqiipubl.com/ojs/index.php/AMS_Journal/article/view/2/2 |
institution |
Universiti Teknologi Malaysia |
building |
UTM Library |
collection |
Institutional Repository |
continent |
Asia |
country |
Malaysia |
content_provider |
Universiti Teknologi Malaysia |
content_source |
UTM Institutional Repository |
url_provider |
http://eprints.utm.my/ |
language |
English |
topic |
TK Electrical engineering. Electronics Nuclear engineering |
spellingShingle |
TK Electrical engineering. Electronics Nuclear engineering Tang, J. W. Hau, Y. W. Marsono, M. N. Hardware/software partitioning of streaming applications for multi-processor system-on-chip |
description |
Hardware/software (HW/SW) co-design has emerged as a crucial and integral part in the development of various embedded applications. Moreover, the increases in the number of embedded multimedia and medical applications make streaming throughput an important attribute of Multi-Processor System-on-Chip (MPSoC). As an important development step, HW/SW partitioning affects the system performance. This paper formulates the optimization of HW/SW partitioning aiming at maximizing streaming throughput with predefined area constraint, targeted for multi-processor system with hardware accelerator sharing capability. Software-oriented and hardware-oriented greedy heuristics for HW/SW partitioning are proposed, as well as a branch-and-bound algorithm with best-first search that utilizes greedy results as initial best solution. Several random graphs and two multimedia applications (JPEG encoder and MP3 decoder) are used for performance benchmarking against brute force ground truth. Results show that the proposed greedy algorithms produce fast solutions which achieve 87.7% and 84.2% near-optimal solution respectively compared to ground truth result. With the aid of greedy result as initial solution, the proposed branch-and-bound algorithm is able to produce ground truth solution up to 2.4741e+8 times faster in HW/SW partitioning time compared to exhaustive brute force method. |
format |
Article |
author |
Tang, J. W. Hau, Y. W. Marsono, M. N. |
author_facet |
Tang, J. W. Hau, Y. W. Marsono, M. N. |
author_sort |
Tang, J. W. |
title |
Hardware/software partitioning of streaming applications for multi-processor system-on-chip |
title_short |
Hardware/software partitioning of streaming applications for multi-processor system-on-chip |
title_full |
Hardware/software partitioning of streaming applications for multi-processor system-on-chip |
title_fullStr |
Hardware/software partitioning of streaming applications for multi-processor system-on-chip |
title_full_unstemmed |
Hardware/software partitioning of streaming applications for multi-processor system-on-chip |
title_sort |
hardware/software partitioning of streaming applications for multi-processor system-on-chip |
publisher |
ARQII Publication, Malaysian Simulation Society |
publishDate |
2017 |
url |
http://eprints.utm.my/id/eprint/80739/1/MuhammadNadzirMarsono2017_HardwareSoftwarePartitioningofStreaming.pdf http://eprints.utm.my/id/eprint/80739/ http://arqiipubl.com/ojs/index.php/AMS_Journal/article/view/2/2 |
_version_ |
1643658502085804032 |
score |
13.15806 |