Machine Learning-Based Processor Adaptability Targeting Energy, Performance, and Reliability

Anderson Luiz Sartor; Pedro Henrique Exenberger Becker; Stephan Wong; Radu Marculescu; Antonio Carlos Schneider Beck

doi:10.1109/ISVLSI.2019.00037

Machine Learning-Based Processor Adaptability Targeting Energy, Performance, and Reliability

Anderson Luiz Sartor, Pedro Henrique Exenberger Becker, Stephan Wong, Radu Marculescu, Antonio Carlos Schneider Beck

Computer Engineering

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

1 Citation (Scopus)

Abstract

Adaptive processors can dynamically change their hardware configuration by tuning several knobs that optimize a given metric, according to the current application. However, the complexity of choosing the best setup at runtime increases exponentially as more adaptive resources become available. Therefore, we propose a polymorphic VLIW processor coupled to a machine learning-based decision mechanism that quickly and accurately delivers the best trade-off in terms of energy, performance, and reliability. The proposed system predicts the best processor configuration in 97.37% of the test cases and achieves an efficiency that is close to an oracle (more than 93.30% on all benchmarks).

Original language	English
Title of host publication	2019 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019
Subtitle of host publication	Proceedings
Editors	L. O'Conner
Place of Publication	Piscataway
Publisher	IEEE
Pages	158-163
Number of pages	6
ISBN (Electronic)	978-1-7281-3391-1
ISBN (Print)	978-1-7281-3392-8
DOIs	https://doi.org/10.1109/ISVLSI.2019.00037
Publication status	Published - 1 Jul 2019
Event	18th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019 - Miami, United States Duration: 15 Jul 2019 → 17 Jul 2019

Conference

Conference	18th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019
Country/Territory	United States
City	Miami
Period	15/07/19 → 17/07/19

Keywords

Adaptive processor
Energy consumption
Fault tolerance
Machine learning
Runtime optimization

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/ISVLSI.2019.00037

Cite this

Sartor, A. L., Exenberger Becker, P. H., Wong, S., Marculescu, R., & Schneider Beck, A. C. (2019). Machine Learning-Based Processor Adaptability Targeting Energy, Performance, and Reliability. In L. O'Conner (Ed.), 2019 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019: Proceedings (pp. 158-163). Article 8839457 IEEE. https://doi.org/10.1109/ISVLSI.2019.00037

@inproceedings{7faf83f9e8ad42e5ab47b8ccb9069124,

title = "Machine Learning-Based Processor Adaptability Targeting Energy, Performance, and Reliability",

abstract = "Adaptive processors can dynamically change their hardware configuration by tuning several knobs that optimize a given metric, according to the current application. However, the complexity of choosing the best setup at runtime increases exponentially as more adaptive resources become available. Therefore, we propose a polymorphic VLIW processor coupled to a machine learning-based decision mechanism that quickly and accurately delivers the best trade-off in terms of energy, performance, and reliability. The proposed system predicts the best processor configuration in 97.37% of the test cases and achieves an efficiency that is close to an oracle (more than 93.30% on all benchmarks).",

keywords = "Adaptive processor, Energy consumption, Fault tolerance, Machine learning, Runtime optimization",

author = "Sartor, {Anderson Luiz} and {Exenberger Becker}, {Pedro Henrique} and Stephan Wong and Radu Marculescu and {Schneider Beck}, {Antonio Carlos}",

year = "2019",

month = jul,

day = "1",

doi = "10.1109/ISVLSI.2019.00037",

language = "English",

isbn = "978-1-7281-3392-8",

pages = "158--163",

editor = "L. O'Conner",

booktitle = "2019 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019",

publisher = "IEEE",

address = "United States",

note = "18th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019 ; Conference date: 15-07-2019 Through 17-07-2019",

}

Sartor, AL, Exenberger Becker, PH, Wong, S, Marculescu, R & Schneider Beck, AC 2019, Machine Learning-Based Processor Adaptability Targeting Energy, Performance, and Reliability. in L O'Conner (ed.), 2019 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019: Proceedings ., 8839457, IEEE, Piscataway, pp. 158-163, 18th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019, Miami, United States, 15/07/19. https://doi.org/10.1109/ISVLSI.2019.00037

Machine Learning-Based Processor Adaptability Targeting Energy, Performance, and Reliability. / Sartor, Anderson Luiz; Exenberger Becker, Pedro Henrique; Wong, Stephan et al.
2019 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019: Proceedings . ed. / L. O'Conner. Piscataway: IEEE, 2019. p. 158-163 8839457.

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Machine Learning-Based Processor Adaptability Targeting Energy, Performance, and Reliability

AU - Sartor, Anderson Luiz

AU - Exenberger Becker, Pedro Henrique

AU - Wong, Stephan

AU - Marculescu, Radu

AU - Schneider Beck, Antonio Carlos

PY - 2019/7/1

Y1 - 2019/7/1

N2 - Adaptive processors can dynamically change their hardware configuration by tuning several knobs that optimize a given metric, according to the current application. However, the complexity of choosing the best setup at runtime increases exponentially as more adaptive resources become available. Therefore, we propose a polymorphic VLIW processor coupled to a machine learning-based decision mechanism that quickly and accurately delivers the best trade-off in terms of energy, performance, and reliability. The proposed system predicts the best processor configuration in 97.37% of the test cases and achieves an efficiency that is close to an oracle (more than 93.30% on all benchmarks).

AB - Adaptive processors can dynamically change their hardware configuration by tuning several knobs that optimize a given metric, according to the current application. However, the complexity of choosing the best setup at runtime increases exponentially as more adaptive resources become available. Therefore, we propose a polymorphic VLIW processor coupled to a machine learning-based decision mechanism that quickly and accurately delivers the best trade-off in terms of energy, performance, and reliability. The proposed system predicts the best processor configuration in 97.37% of the test cases and achieves an efficiency that is close to an oracle (more than 93.30% on all benchmarks).

KW - Adaptive processor

KW - Energy consumption

KW - Fault tolerance

KW - Machine learning

KW - Runtime optimization

UR - http://www.scopus.com/inward/record.url?scp=85072986589&partnerID=8YFLogxK

U2 - 10.1109/ISVLSI.2019.00037

DO - 10.1109/ISVLSI.2019.00037

M3 - Conference contribution

AN - SCOPUS:85072986589

SN - 978-1-7281-3392-8

SP - 158

EP - 163

BT - 2019 IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019

A2 - O'Conner, L.

PB - IEEE

CY - Piscataway

T2 - 18th IEEE Computer Society Annual Symposium on VLSI, ISVLSI 2019

Y2 - 15 July 2019 through 17 July 2019

ER -

Machine Learning-Based Processor Adaptability Targeting Energy, Performance, and Reliability

Abstract

Conference

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this