Non-intrusive dynamic application profiling for multitasked applications

Karthik Shankar, Roman L Lysecky

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

9 Citations (Scopus)

Abstract

Application profiling - the process of monitoring an application to determine the frequency of execution within specific regions - is an essential step within the design process for many software and hardware systems. Profiling is often a critical step within hardware/software partitioning utilized to determine the critical kernels of an application. In this paper, we present a non-intrusive dynamic application profiler (DAProf) capable of profiling an executing application by monitoring the application's short backwards branches, function calls, function returns, as well as efficiently detecting context switches to provide accurate characterization of the frequently executed loops within multitasked applications. DAProf can accurately profile multiple tasks within a software application with 98.5% accuracy using as little as 10% additional area compared to an ARM9 processor.

Original languageEnglish (US)
Title of host publicationProceedings - Design Automation Conference
Pages130-135
Number of pages6
StatePublished - 2009
Event2009 46th ACM/IEEE Design Automation Conference, DAC 2009 - San Francisco, CA, United States
Duration: Jul 26 2009Jul 31 2009

Other

Other2009 46th ACM/IEEE Design Automation Conference, DAC 2009
CountryUnited States
CitySan Francisco, CA
Period7/26/097/31/09

Fingerprint

Profiling
Hardware/software Partitioning
Hardware
Monitoring
Software
Application programs
Design Process
Switch
Switches
Branch
kernel

Keywords

  • Dynamic hardware/software partitioning
  • Dynamic optimizations
  • Multitasking
  • Profiling
  • Real-time embedded systems

ASJC Scopus subject areas

  • Computer Science Applications
  • Control and Systems Engineering
  • Electrical and Electronic Engineering
  • Modeling and Simulation

Cite this

Shankar, K., & Lysecky, R. L. (2009). Non-intrusive dynamic application profiling for multitasked applications. In Proceedings - Design Automation Conference (pp. 130-135). [5227185]

Non-intrusive dynamic application profiling for multitasked applications. / Shankar, Karthik; Lysecky, Roman L.

Proceedings - Design Automation Conference. 2009. p. 130-135 5227185.

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

Shankar, K & Lysecky, RL 2009, Non-intrusive dynamic application profiling for multitasked applications. in Proceedings - Design Automation Conference., 5227185, pp. 130-135, 2009 46th ACM/IEEE Design Automation Conference, DAC 2009, San Francisco, CA, United States, 7/26/09.
Shankar K, Lysecky RL. Non-intrusive dynamic application profiling for multitasked applications. In Proceedings - Design Automation Conference. 2009. p. 130-135. 5227185
Shankar, Karthik ; Lysecky, Roman L. / Non-intrusive dynamic application profiling for multitasked applications. Proceedings - Design Automation Conference. 2009. pp. 130-135
@inproceedings{7072faf2140a4f2b8aa8228fc470533d,
title = "Non-intrusive dynamic application profiling for multitasked applications",
abstract = "Application profiling - the process of monitoring an application to determine the frequency of execution within specific regions - is an essential step within the design process for many software and hardware systems. Profiling is often a critical step within hardware/software partitioning utilized to determine the critical kernels of an application. In this paper, we present a non-intrusive dynamic application profiler (DAProf) capable of profiling an executing application by monitoring the application's short backwards branches, function calls, function returns, as well as efficiently detecting context switches to provide accurate characterization of the frequently executed loops within multitasked applications. DAProf can accurately profile multiple tasks within a software application with 98.5{\%} accuracy using as little as 10{\%} additional area compared to an ARM9 processor.",
keywords = "Dynamic hardware/software partitioning, Dynamic optimizations, Multitasking, Profiling, Real-time embedded systems",
author = "Karthik Shankar and Lysecky, {Roman L}",
year = "2009",
language = "English (US)",
isbn = "9781605584973",
pages = "130--135",
booktitle = "Proceedings - Design Automation Conference",

}

TY - GEN

T1 - Non-intrusive dynamic application profiling for multitasked applications

AU - Shankar, Karthik

AU - Lysecky, Roman L

PY - 2009

Y1 - 2009

N2 - Application profiling - the process of monitoring an application to determine the frequency of execution within specific regions - is an essential step within the design process for many software and hardware systems. Profiling is often a critical step within hardware/software partitioning utilized to determine the critical kernels of an application. In this paper, we present a non-intrusive dynamic application profiler (DAProf) capable of profiling an executing application by monitoring the application's short backwards branches, function calls, function returns, as well as efficiently detecting context switches to provide accurate characterization of the frequently executed loops within multitasked applications. DAProf can accurately profile multiple tasks within a software application with 98.5% accuracy using as little as 10% additional area compared to an ARM9 processor.

AB - Application profiling - the process of monitoring an application to determine the frequency of execution within specific regions - is an essential step within the design process for many software and hardware systems. Profiling is often a critical step within hardware/software partitioning utilized to determine the critical kernels of an application. In this paper, we present a non-intrusive dynamic application profiler (DAProf) capable of profiling an executing application by monitoring the application's short backwards branches, function calls, function returns, as well as efficiently detecting context switches to provide accurate characterization of the frequently executed loops within multitasked applications. DAProf can accurately profile multiple tasks within a software application with 98.5% accuracy using as little as 10% additional area compared to an ARM9 processor.

KW - Dynamic hardware/software partitioning

KW - Dynamic optimizations

KW - Multitasking

KW - Profiling

KW - Real-time embedded systems

UR - http://www.scopus.com/inward/record.url?scp=70350719483&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70350719483&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:70350719483

SN - 9781605584973

SP - 130

EP - 135

BT - Proceedings - Design Automation Conference

ER -