An effective methodology to multi-objective design of application domain-specific embedded architectures

Vincenzo Catania, Alessandro G. Di Nuovo, Maurizio Palesi, Davide Patti, Gianmarco Morales

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Today's computer systems have become unbelievably complex. Nowadays register-level design is an overwhelming task, especially in the embedded system area where the time-to-market is very short. Platform Based Design shifts the challenge on how to tune parametric platforms to achieve the best performance at the smallest cost. This task, called multi-objective Design Space Exploration, requires accurate strategies because the design space is too vast to be exhaustively evaluated. Even using efficient exploration strategies proposed in the literature, simulation times can become a bottleneck in the design flow. In this work we propose a novel approach to application-domain Design Space Exploration using a Multi-Objective Genetic Algorithm and employing HPC to reduce exploration times. The genetic algorithm is preceded by a correlation analysis of the different objectives. The search space is thus reduced by combining highly correlated objectives from different domains. We describe the steps needed to parallelize the exploration on the Grid, and present the results of extensive testing of the proposed approach. We obtained over one order of magnitude reduction in exploration times without hampering the quality of the solutions. Shorter simulation times allow more ideas to be explored in less time. This leads to shorter product time-to-market and a more thorough design space exploration. Furthermore the combination of correlated objectives favors the design of modern multi-purpose devices.

Original languageEnglish
Title of host publication12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009
Pages643-650
Number of pages8
DOIs
Publication statusPublished - 2009
Externally publishedYes
Event12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009 - Patras, Greece
Duration: 27 Aug 200929 Aug 2009

Other

Other12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009
CountryGreece
CityPatras
Period27/8/0929/8/09

Fingerprint

Genetic algorithms
Embedded systems
Computer systems
Testing
Costs

ASJC Scopus subject areas

  • Hardware and Architecture
  • Software
  • Electrical and Electronic Engineering

Cite this

Catania, V., Di Nuovo, A. G., Palesi, M., Patti, D., & Morales, G. (2009). An effective methodology to multi-objective design of application domain-specific embedded architectures. In 12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009 (pp. 643-650). [5350191] https://doi.org/10.1109/DSD.2009.205

An effective methodology to multi-objective design of application domain-specific embedded architectures. / Catania, Vincenzo; Di Nuovo, Alessandro G.; Palesi, Maurizio; Patti, Davide; Morales, Gianmarco.

12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009. 2009. p. 643-650 5350191.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Catania, V, Di Nuovo, AG, Palesi, M, Patti, D & Morales, G 2009, An effective methodology to multi-objective design of application domain-specific embedded architectures. in 12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009., 5350191, pp. 643-650, 12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009, Patras, Greece, 27/8/09. https://doi.org/10.1109/DSD.2009.205
Catania V, Di Nuovo AG, Palesi M, Patti D, Morales G. An effective methodology to multi-objective design of application domain-specific embedded architectures. In 12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009. 2009. p. 643-650. 5350191 https://doi.org/10.1109/DSD.2009.205
Catania, Vincenzo ; Di Nuovo, Alessandro G. ; Palesi, Maurizio ; Patti, Davide ; Morales, Gianmarco. / An effective methodology to multi-objective design of application domain-specific embedded architectures. 12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009. 2009. pp. 643-650
@inproceedings{2d4c5d7e094a4574a1b5bda4b22d489f,
title = "An effective methodology to multi-objective design of application domain-specific embedded architectures",
abstract = "Today's computer systems have become unbelievably complex. Nowadays register-level design is an overwhelming task, especially in the embedded system area where the time-to-market is very short. Platform Based Design shifts the challenge on how to tune parametric platforms to achieve the best performance at the smallest cost. This task, called multi-objective Design Space Exploration, requires accurate strategies because the design space is too vast to be exhaustively evaluated. Even using efficient exploration strategies proposed in the literature, simulation times can become a bottleneck in the design flow. In this work we propose a novel approach to application-domain Design Space Exploration using a Multi-Objective Genetic Algorithm and employing HPC to reduce exploration times. The genetic algorithm is preceded by a correlation analysis of the different objectives. The search space is thus reduced by combining highly correlated objectives from different domains. We describe the steps needed to parallelize the exploration on the Grid, and present the results of extensive testing of the proposed approach. We obtained over one order of magnitude reduction in exploration times without hampering the quality of the solutions. Shorter simulation times allow more ideas to be explored in less time. This leads to shorter product time-to-market and a more thorough design space exploration. Furthermore the combination of correlated objectives favors the design of modern multi-purpose devices.",
author = "Vincenzo Catania and {Di Nuovo}, {Alessandro G.} and Maurizio Palesi and Davide Patti and Gianmarco Morales",
year = "2009",
doi = "10.1109/DSD.2009.205",
language = "English",
isbn = "9780769537825",
pages = "643--650",
booktitle = "12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009",

}

TY - GEN

T1 - An effective methodology to multi-objective design of application domain-specific embedded architectures

AU - Catania, Vincenzo

AU - Di Nuovo, Alessandro G.

AU - Palesi, Maurizio

AU - Patti, Davide

AU - Morales, Gianmarco

PY - 2009

Y1 - 2009

N2 - Today's computer systems have become unbelievably complex. Nowadays register-level design is an overwhelming task, especially in the embedded system area where the time-to-market is very short. Platform Based Design shifts the challenge on how to tune parametric platforms to achieve the best performance at the smallest cost. This task, called multi-objective Design Space Exploration, requires accurate strategies because the design space is too vast to be exhaustively evaluated. Even using efficient exploration strategies proposed in the literature, simulation times can become a bottleneck in the design flow. In this work we propose a novel approach to application-domain Design Space Exploration using a Multi-Objective Genetic Algorithm and employing HPC to reduce exploration times. The genetic algorithm is preceded by a correlation analysis of the different objectives. The search space is thus reduced by combining highly correlated objectives from different domains. We describe the steps needed to parallelize the exploration on the Grid, and present the results of extensive testing of the proposed approach. We obtained over one order of magnitude reduction in exploration times without hampering the quality of the solutions. Shorter simulation times allow more ideas to be explored in less time. This leads to shorter product time-to-market and a more thorough design space exploration. Furthermore the combination of correlated objectives favors the design of modern multi-purpose devices.

AB - Today's computer systems have become unbelievably complex. Nowadays register-level design is an overwhelming task, especially in the embedded system area where the time-to-market is very short. Platform Based Design shifts the challenge on how to tune parametric platforms to achieve the best performance at the smallest cost. This task, called multi-objective Design Space Exploration, requires accurate strategies because the design space is too vast to be exhaustively evaluated. Even using efficient exploration strategies proposed in the literature, simulation times can become a bottleneck in the design flow. In this work we propose a novel approach to application-domain Design Space Exploration using a Multi-Objective Genetic Algorithm and employing HPC to reduce exploration times. The genetic algorithm is preceded by a correlation analysis of the different objectives. The search space is thus reduced by combining highly correlated objectives from different domains. We describe the steps needed to parallelize the exploration on the Grid, and present the results of extensive testing of the proposed approach. We obtained over one order of magnitude reduction in exploration times without hampering the quality of the solutions. Shorter simulation times allow more ideas to be explored in less time. This leads to shorter product time-to-market and a more thorough design space exploration. Furthermore the combination of correlated objectives favors the design of modern multi-purpose devices.

UR - http://www.scopus.com/inward/record.url?scp=74549142413&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=74549142413&partnerID=8YFLogxK

U2 - 10.1109/DSD.2009.205

DO - 10.1109/DSD.2009.205

M3 - Conference contribution

SN - 9780769537825

SP - 643

EP - 650

BT - 12th Euromicro Conference on Digital System Design: Architectures, Methods and Tools, DSD 2009

ER -