Spendthrift: Machine learning based resource and frequency scaling for ambient energy harvesting nonvolatile processors

Kaisheng Ma, Xueqing Li, Srivatsa Rangachar Srinivasa, Yongpan Liu, John Morgan Sampson, Yuan Xie, Vijaykrishnan Narayanan

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

14 Citations (Scopus)

Abstract

Batteryless energy harvesting systems face a twofold challenge in converting incoming energy into forward progress. Not only must such systems contend with inherently weak and fluctuating power sources, but they have very limited temporal windows for capitalizing on transitory periods of above-average power. To maximize forward progress, such systems should aggressively consume energy when it is available, rather than optimizing for peak averagecase efficiency. However, there are multiple ways that a processor can trade between consumption and performance. In this paper, we examine two approaches, frequency scaling and resource scaling, and develop a predictor-driven scheme for dynamically allocating future power budgets between the two techniques. We show that our solution can achieve forward progress equal to 2.08X of the baseline Out-of-Order (OoO) processor with the best static configuration of frequency and resources. The combined technique outperforms either technique in isolation, with frequency-only and resource-only approaches achieving 1.43X and 1.61X forward progress improvements, respectively.

Original languageEnglish (US)
Title of host publication2017 22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages678-683
Number of pages6
ISBN (Electronic)9781509015580
DOIs
StatePublished - Feb 16 2017
Event22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017 - Chiba, Japan
Duration: Jan 16 2017Jan 19 2017

Publication series

NameProceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC

Other

Other22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017
CountryJapan
CityChiba
Period1/16/171/19/17

Fingerprint

Energy harvesting
Learning systems

All Science Journal Classification (ASJC) codes

  • Electrical and Electronic Engineering
  • Computer Science Applications
  • Computer Graphics and Computer-Aided Design

Cite this

Ma, K., Li, X., Srinivasa, S. R., Liu, Y., Sampson, J. M., Xie, Y., & Narayanan, V. (2017). Spendthrift: Machine learning based resource and frequency scaling for ambient energy harvesting nonvolatile processors. In 2017 22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017 (pp. 678-683). [7858402] (Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ASPDAC.2017.7858402
Ma, Kaisheng ; Li, Xueqing ; Srinivasa, Srivatsa Rangachar ; Liu, Yongpan ; Sampson, John Morgan ; Xie, Yuan ; Narayanan, Vijaykrishnan. / Spendthrift : Machine learning based resource and frequency scaling for ambient energy harvesting nonvolatile processors. 2017 22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 678-683 (Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC).
@inproceedings{210cd699823e436dace5f457df5f3923,
title = "Spendthrift: Machine learning based resource and frequency scaling for ambient energy harvesting nonvolatile processors",
abstract = "Batteryless energy harvesting systems face a twofold challenge in converting incoming energy into forward progress. Not only must such systems contend with inherently weak and fluctuating power sources, but they have very limited temporal windows for capitalizing on transitory periods of above-average power. To maximize forward progress, such systems should aggressively consume energy when it is available, rather than optimizing for peak averagecase efficiency. However, there are multiple ways that a processor can trade between consumption and performance. In this paper, we examine two approaches, frequency scaling and resource scaling, and develop a predictor-driven scheme for dynamically allocating future power budgets between the two techniques. We show that our solution can achieve forward progress equal to 2.08X of the baseline Out-of-Order (OoO) processor with the best static configuration of frequency and resources. The combined technique outperforms either technique in isolation, with frequency-only and resource-only approaches achieving 1.43X and 1.61X forward progress improvements, respectively.",
author = "Kaisheng Ma and Xueqing Li and Srinivasa, {Srivatsa Rangachar} and Yongpan Liu and Sampson, {John Morgan} and Yuan Xie and Vijaykrishnan Narayanan",
year = "2017",
month = "2",
day = "16",
doi = "10.1109/ASPDAC.2017.7858402",
language = "English (US)",
series = "Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "678--683",
booktitle = "2017 22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017",
address = "United States",

}

Ma, K, Li, X, Srinivasa, SR, Liu, Y, Sampson, JM, Xie, Y & Narayanan, V 2017, Spendthrift: Machine learning based resource and frequency scaling for ambient energy harvesting nonvolatile processors. in 2017 22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017., 7858402, Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC, Institute of Electrical and Electronics Engineers Inc., pp. 678-683, 22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017, Chiba, Japan, 1/16/17. https://doi.org/10.1109/ASPDAC.2017.7858402

Spendthrift : Machine learning based resource and frequency scaling for ambient energy harvesting nonvolatile processors. / Ma, Kaisheng; Li, Xueqing; Srinivasa, Srivatsa Rangachar; Liu, Yongpan; Sampson, John Morgan; Xie, Yuan; Narayanan, Vijaykrishnan.

2017 22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017. Institute of Electrical and Electronics Engineers Inc., 2017. p. 678-683 7858402 (Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC).

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

TY - GEN

T1 - Spendthrift

T2 - Machine learning based resource and frequency scaling for ambient energy harvesting nonvolatile processors

AU - Ma, Kaisheng

AU - Li, Xueqing

AU - Srinivasa, Srivatsa Rangachar

AU - Liu, Yongpan

AU - Sampson, John Morgan

AU - Xie, Yuan

AU - Narayanan, Vijaykrishnan

PY - 2017/2/16

Y1 - 2017/2/16

N2 - Batteryless energy harvesting systems face a twofold challenge in converting incoming energy into forward progress. Not only must such systems contend with inherently weak and fluctuating power sources, but they have very limited temporal windows for capitalizing on transitory periods of above-average power. To maximize forward progress, such systems should aggressively consume energy when it is available, rather than optimizing for peak averagecase efficiency. However, there are multiple ways that a processor can trade between consumption and performance. In this paper, we examine two approaches, frequency scaling and resource scaling, and develop a predictor-driven scheme for dynamically allocating future power budgets between the two techniques. We show that our solution can achieve forward progress equal to 2.08X of the baseline Out-of-Order (OoO) processor with the best static configuration of frequency and resources. The combined technique outperforms either technique in isolation, with frequency-only and resource-only approaches achieving 1.43X and 1.61X forward progress improvements, respectively.

AB - Batteryless energy harvesting systems face a twofold challenge in converting incoming energy into forward progress. Not only must such systems contend with inherently weak and fluctuating power sources, but they have very limited temporal windows for capitalizing on transitory periods of above-average power. To maximize forward progress, such systems should aggressively consume energy when it is available, rather than optimizing for peak averagecase efficiency. However, there are multiple ways that a processor can trade between consumption and performance. In this paper, we examine two approaches, frequency scaling and resource scaling, and develop a predictor-driven scheme for dynamically allocating future power budgets between the two techniques. We show that our solution can achieve forward progress equal to 2.08X of the baseline Out-of-Order (OoO) processor with the best static configuration of frequency and resources. The combined technique outperforms either technique in isolation, with frequency-only and resource-only approaches achieving 1.43X and 1.61X forward progress improvements, respectively.

UR - http://www.scopus.com/inward/record.url?scp=85015262484&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015262484&partnerID=8YFLogxK

U2 - 10.1109/ASPDAC.2017.7858402

DO - 10.1109/ASPDAC.2017.7858402

M3 - Conference contribution

AN - SCOPUS:85015262484

T3 - Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC

SP - 678

EP - 683

BT - 2017 22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Ma K, Li X, Srinivasa SR, Liu Y, Sampson JM, Xie Y et al. Spendthrift: Machine learning based resource and frequency scaling for ambient energy harvesting nonvolatile processors. In 2017 22nd Asia and South Pacific Design Automation Conference, ASP-DAC 2017. Institute of Electrical and Electronics Engineers Inc. 2017. p. 678-683. 7858402. (Proceedings of the Asia and South Pacific Design Automation Conference, ASP-DAC). https://doi.org/10.1109/ASPDAC.2017.7858402