Hybrid techniques for fast multicore simulation

Manu Shantharam, Padma Raghavan, Mahmut Kandemir

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

Abstract

One of the challenges in the design of multicore architectures concerns the fast evaluation of hardware design-tradeoffs using simulation techniques. Simulation tools for multicore architectures tend to have long execution times that grow linearly with the number of cores simulated. In this paper, we present two hybrid techniques for fast and accurate multicore simulation. Our first method, the Monte Carlo Co-Simulation (MCCS) scheme, considers application phases, and within each phase, interleaves a Monte Carlo modeling scheme with a traditional simulator, such as Simics. Our second method, the Curve Fitting Based Simulation (CFBS) scheme, is tailored to evaluate the behavior of applications with multiple iterations, such as scientific applications that have consistent cycles per instruction (CPI) behavior within a subroutine over different iterations. In our CFBS method, we represent the CPI profile of a subroutine as a signature using curve fitting and represent the entire application execution as a set of signatures to predict performance metrics. Our results indicate that MCCS can reduce simulation time by as much as a factor of 2.37, with a speedup of 1.77 on average compared to Simics. We also observe that CFBS can reduce simulation time by as much as a factor of 13.6, with a speedup of 6.24 on average. The observed average relative errors in CPI compared to Simics are 32% for MCCS and significantly lower, at 2%, for CFBS.

Original languageEnglish (US)
Title of host publicationEuro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings
Pages122-134
Number of pages13
DOIs
StatePublished - Nov 9 2009
EventEuro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings - Delft, Netherlands
Duration: Aug 25 2009Aug 28 2009

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5704 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

OtherEuro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings
CountryNetherlands
CityDelft
Period8/25/098/28/09

Fingerprint

Curve fitting
Co-simulation
Subroutines
Simulation
Cycle
Speedup
Signature
Iteration
Hardware Design
Simulators
Performance Metrics
Relative Error
Simulation Tool
Simulation Methods
Hardware
Execution Time
Simulator
Linearly
Trade-offs
Entire

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Shantharam, M., Raghavan, P., & Kandemir, M. (2009). Hybrid techniques for fast multicore simulation. In Euro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings (pp. 122-134). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5704 LNCS). https://doi.org/10.1007/978-3-642-03869-3_15
Shantharam, Manu ; Raghavan, Padma ; Kandemir, Mahmut. / Hybrid techniques for fast multicore simulation. Euro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings. 2009. pp. 122-134 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{c7066e3ac1c946f18bf957679998edad,
title = "Hybrid techniques for fast multicore simulation",
abstract = "One of the challenges in the design of multicore architectures concerns the fast evaluation of hardware design-tradeoffs using simulation techniques. Simulation tools for multicore architectures tend to have long execution times that grow linearly with the number of cores simulated. In this paper, we present two hybrid techniques for fast and accurate multicore simulation. Our first method, the Monte Carlo Co-Simulation (MCCS) scheme, considers application phases, and within each phase, interleaves a Monte Carlo modeling scheme with a traditional simulator, such as Simics. Our second method, the Curve Fitting Based Simulation (CFBS) scheme, is tailored to evaluate the behavior of applications with multiple iterations, such as scientific applications that have consistent cycles per instruction (CPI) behavior within a subroutine over different iterations. In our CFBS method, we represent the CPI profile of a subroutine as a signature using curve fitting and represent the entire application execution as a set of signatures to predict performance metrics. Our results indicate that MCCS can reduce simulation time by as much as a factor of 2.37, with a speedup of 1.77 on average compared to Simics. We also observe that CFBS can reduce simulation time by as much as a factor of 13.6, with a speedup of 6.24 on average. The observed average relative errors in CPI compared to Simics are 32{\%} for MCCS and significantly lower, at 2{\%}, for CFBS.",
author = "Manu Shantharam and Padma Raghavan and Mahmut Kandemir",
year = "2009",
month = "11",
day = "9",
doi = "10.1007/978-3-642-03869-3_15",
language = "English (US)",
isbn = "3642038689",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "122--134",
booktitle = "Euro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings",

}

Shantharam, M, Raghavan, P & Kandemir, M 2009, Hybrid techniques for fast multicore simulation. in Euro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5704 LNCS, pp. 122-134, Euro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings, Delft, Netherlands, 8/25/09. https://doi.org/10.1007/978-3-642-03869-3_15

Hybrid techniques for fast multicore simulation. / Shantharam, Manu; Raghavan, Padma; Kandemir, Mahmut.

Euro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings. 2009. p. 122-134 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5704 LNCS).

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

TY - GEN

T1 - Hybrid techniques for fast multicore simulation

AU - Shantharam, Manu

AU - Raghavan, Padma

AU - Kandemir, Mahmut

PY - 2009/11/9

Y1 - 2009/11/9

N2 - One of the challenges in the design of multicore architectures concerns the fast evaluation of hardware design-tradeoffs using simulation techniques. Simulation tools for multicore architectures tend to have long execution times that grow linearly with the number of cores simulated. In this paper, we present two hybrid techniques for fast and accurate multicore simulation. Our first method, the Monte Carlo Co-Simulation (MCCS) scheme, considers application phases, and within each phase, interleaves a Monte Carlo modeling scheme with a traditional simulator, such as Simics. Our second method, the Curve Fitting Based Simulation (CFBS) scheme, is tailored to evaluate the behavior of applications with multiple iterations, such as scientific applications that have consistent cycles per instruction (CPI) behavior within a subroutine over different iterations. In our CFBS method, we represent the CPI profile of a subroutine as a signature using curve fitting and represent the entire application execution as a set of signatures to predict performance metrics. Our results indicate that MCCS can reduce simulation time by as much as a factor of 2.37, with a speedup of 1.77 on average compared to Simics. We also observe that CFBS can reduce simulation time by as much as a factor of 13.6, with a speedup of 6.24 on average. The observed average relative errors in CPI compared to Simics are 32% for MCCS and significantly lower, at 2%, for CFBS.

AB - One of the challenges in the design of multicore architectures concerns the fast evaluation of hardware design-tradeoffs using simulation techniques. Simulation tools for multicore architectures tend to have long execution times that grow linearly with the number of cores simulated. In this paper, we present two hybrid techniques for fast and accurate multicore simulation. Our first method, the Monte Carlo Co-Simulation (MCCS) scheme, considers application phases, and within each phase, interleaves a Monte Carlo modeling scheme with a traditional simulator, such as Simics. Our second method, the Curve Fitting Based Simulation (CFBS) scheme, is tailored to evaluate the behavior of applications with multiple iterations, such as scientific applications that have consistent cycles per instruction (CPI) behavior within a subroutine over different iterations. In our CFBS method, we represent the CPI profile of a subroutine as a signature using curve fitting and represent the entire application execution as a set of signatures to predict performance metrics. Our results indicate that MCCS can reduce simulation time by as much as a factor of 2.37, with a speedup of 1.77 on average compared to Simics. We also observe that CFBS can reduce simulation time by as much as a factor of 13.6, with a speedup of 6.24 on average. The observed average relative errors in CPI compared to Simics are 32% for MCCS and significantly lower, at 2%, for CFBS.

UR - http://www.scopus.com/inward/record.url?scp=70350627376&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70350627376&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-03869-3_15

DO - 10.1007/978-3-642-03869-3_15

M3 - Conference contribution

AN - SCOPUS:70350627376

SN - 3642038689

SN - 9783642038686

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 122

EP - 134

BT - Euro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings

ER -

Shantharam M, Raghavan P, Kandemir M. Hybrid techniques for fast multicore simulation. In Euro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings. 2009. p. 122-134. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-03869-3_15