A case for temperature-driven task migration to balance energy efficiency and image quality of vision processing workloads

Venkatesh Kodukula, Sai Bharadwaj Medapuram, Britton Jones, Robert LiKamWa

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

1 Citation (Scopus)

Abstract

Many researchers in academia and industry [4, 8] advocate shifting processing near the image sensor through near-sensor accelerators to reduce data movement across energy-expensive interfaces. However, near-sensor processing also heats the sensor, increasing thermal noise and hot pixels, which degrades image quality. To understand these implications, we perform an energy and thermal characterization in the context of an augmented reality case study around visual marker detection. Our characterization results show that for a near-sensor accelerator consuming 1 W of power, dynamic range drops by 16 dB, image noise increases by 3 times, and the number of hot pixels multiplies by 16, degrading image quality. Such degradation impairs the task accuracy of interactive perceptual applications that require high accuracy. The marker-detection fails for 12% of frames when degraded by 1 minute of 1 W near-sensor power consumption. To this end, we propose temperature-driven task migration, a system-level technique that partitions processing between the thermally-coupled near-sensor accelerator and the thermally-isolated CPU host. Leveraging the sensor’s current temperature and application driven image fidelity requirements, this technique mitigates task accuracy issues while providing gains in energy-efficiency. We discuss challenges pertaining to effective, seamless migration decisions at runtime, and propose potential solutions.

Original languageEnglish (US)
Title of host publicationHotMobile 2018 - Proceedings of the 19th International Workshop on Mobile Computing Systems and Applications
PublisherAssociation for Computing Machinery, Inc
Pages93-98
Number of pages6
Volume2018-February
ISBN (Electronic)9781450356305
DOIs
StatePublished - Feb 12 2018
Event19th International Workshop on Mobile Computing Systems and Applications, HotMobile 2018 - Tempe, United States
Duration: Feb 12 2018Feb 13 2018

Other

Other19th International Workshop on Mobile Computing Systems and Applications, HotMobile 2018
CountryUnited States
CityTempe
Period2/12/182/13/18

Fingerprint

Image quality
Energy efficiency
Sensors
Processing
Particle accelerators
Temperature
Pixels
Thermal noise
Augmented reality
Image sensors
Program processors
Electric power utilization
Degradation
Industry

Keywords

  • Continuous sensing
  • Mobile systems
  • Task migration

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Computer Science Applications
  • Software
  • Computer Networks and Communications

Cite this

Kodukula, V., Bharadwaj Medapuram, S., Jones, B., & LiKamWa, R. (2018). A case for temperature-driven task migration to balance energy efficiency and image quality of vision processing workloads. In HotMobile 2018 - Proceedings of the 19th International Workshop on Mobile Computing Systems and Applications (Vol. 2018-February, pp. 93-98). Association for Computing Machinery, Inc. https://doi.org/10.1145/3177102.3177111

A case for temperature-driven task migration to balance energy efficiency and image quality of vision processing workloads. / Kodukula, Venkatesh; Bharadwaj Medapuram, Sai; Jones, Britton; LiKamWa, Robert.

HotMobile 2018 - Proceedings of the 19th International Workshop on Mobile Computing Systems and Applications. Vol. 2018-February Association for Computing Machinery, Inc, 2018. p. 93-98.

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

Kodukula, V, Bharadwaj Medapuram, S, Jones, B & LiKamWa, R 2018, A case for temperature-driven task migration to balance energy efficiency and image quality of vision processing workloads. in HotMobile 2018 - Proceedings of the 19th International Workshop on Mobile Computing Systems and Applications. vol. 2018-February, Association for Computing Machinery, Inc, pp. 93-98, 19th International Workshop on Mobile Computing Systems and Applications, HotMobile 2018, Tempe, United States, 2/12/18. https://doi.org/10.1145/3177102.3177111
Kodukula V, Bharadwaj Medapuram S, Jones B, LiKamWa R. A case for temperature-driven task migration to balance energy efficiency and image quality of vision processing workloads. In HotMobile 2018 - Proceedings of the 19th International Workshop on Mobile Computing Systems and Applications. Vol. 2018-February. Association for Computing Machinery, Inc. 2018. p. 93-98 https://doi.org/10.1145/3177102.3177111
Kodukula, Venkatesh ; Bharadwaj Medapuram, Sai ; Jones, Britton ; LiKamWa, Robert. / A case for temperature-driven task migration to balance energy efficiency and image quality of vision processing workloads. HotMobile 2018 - Proceedings of the 19th International Workshop on Mobile Computing Systems and Applications. Vol. 2018-February Association for Computing Machinery, Inc, 2018. pp. 93-98
@inproceedings{ed8c6175dc1a4a64b4755daf4faebe48,
title = "A case for temperature-driven task migration to balance energy efficiency and image quality of vision processing workloads",
abstract = "Many researchers in academia and industry [4, 8] advocate shifting processing near the image sensor through near-sensor accelerators to reduce data movement across energy-expensive interfaces. However, near-sensor processing also heats the sensor, increasing thermal noise and hot pixels, which degrades image quality. To understand these implications, we perform an energy and thermal characterization in the context of an augmented reality case study around visual marker detection. Our characterization results show that for a near-sensor accelerator consuming 1 W of power, dynamic range drops by 16 dB, image noise increases by 3 times, and the number of hot pixels multiplies by 16, degrading image quality. Such degradation impairs the task accuracy of interactive perceptual applications that require high accuracy. The marker-detection fails for 12{\%} of frames when degraded by 1 minute of 1 W near-sensor power consumption. To this end, we propose temperature-driven task migration, a system-level technique that partitions processing between the thermally-coupled near-sensor accelerator and the thermally-isolated CPU host. Leveraging the sensor’s current temperature and application driven image fidelity requirements, this technique mitigates task accuracy issues while providing gains in energy-efficiency. We discuss challenges pertaining to effective, seamless migration decisions at runtime, and propose potential solutions.",
keywords = "Continuous sensing, Mobile systems, Task migration",
author = "Venkatesh Kodukula and {Bharadwaj Medapuram}, Sai and Britton Jones and Robert LiKamWa",
year = "2018",
month = "2",
day = "12",
doi = "10.1145/3177102.3177111",
language = "English (US)",
volume = "2018-February",
pages = "93--98",
booktitle = "HotMobile 2018 - Proceedings of the 19th International Workshop on Mobile Computing Systems and Applications",
publisher = "Association for Computing Machinery, Inc",

}

TY - GEN

T1 - A case for temperature-driven task migration to balance energy efficiency and image quality of vision processing workloads

AU - Kodukula, Venkatesh

AU - Bharadwaj Medapuram, Sai

AU - Jones, Britton

AU - LiKamWa, Robert

PY - 2018/2/12

Y1 - 2018/2/12

N2 - Many researchers in academia and industry [4, 8] advocate shifting processing near the image sensor through near-sensor accelerators to reduce data movement across energy-expensive interfaces. However, near-sensor processing also heats the sensor, increasing thermal noise and hot pixels, which degrades image quality. To understand these implications, we perform an energy and thermal characterization in the context of an augmented reality case study around visual marker detection. Our characterization results show that for a near-sensor accelerator consuming 1 W of power, dynamic range drops by 16 dB, image noise increases by 3 times, and the number of hot pixels multiplies by 16, degrading image quality. Such degradation impairs the task accuracy of interactive perceptual applications that require high accuracy. The marker-detection fails for 12% of frames when degraded by 1 minute of 1 W near-sensor power consumption. To this end, we propose temperature-driven task migration, a system-level technique that partitions processing between the thermally-coupled near-sensor accelerator and the thermally-isolated CPU host. Leveraging the sensor’s current temperature and application driven image fidelity requirements, this technique mitigates task accuracy issues while providing gains in energy-efficiency. We discuss challenges pertaining to effective, seamless migration decisions at runtime, and propose potential solutions.

AB - Many researchers in academia and industry [4, 8] advocate shifting processing near the image sensor through near-sensor accelerators to reduce data movement across energy-expensive interfaces. However, near-sensor processing also heats the sensor, increasing thermal noise and hot pixels, which degrades image quality. To understand these implications, we perform an energy and thermal characterization in the context of an augmented reality case study around visual marker detection. Our characterization results show that for a near-sensor accelerator consuming 1 W of power, dynamic range drops by 16 dB, image noise increases by 3 times, and the number of hot pixels multiplies by 16, degrading image quality. Such degradation impairs the task accuracy of interactive perceptual applications that require high accuracy. The marker-detection fails for 12% of frames when degraded by 1 minute of 1 W near-sensor power consumption. To this end, we propose temperature-driven task migration, a system-level technique that partitions processing between the thermally-coupled near-sensor accelerator and the thermally-isolated CPU host. Leveraging the sensor’s current temperature and application driven image fidelity requirements, this technique mitigates task accuracy issues while providing gains in energy-efficiency. We discuss challenges pertaining to effective, seamless migration decisions at runtime, and propose potential solutions.

KW - Continuous sensing

KW - Mobile systems

KW - Task migration

UR - http://www.scopus.com/inward/record.url?scp=85048529280&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048529280&partnerID=8YFLogxK

U2 - 10.1145/3177102.3177111

DO - 10.1145/3177102.3177111

M3 - Conference contribution

AN - SCOPUS:85048529280

VL - 2018-February

SP - 93

EP - 98

BT - HotMobile 2018 - Proceedings of the 19th International Workshop on Mobile Computing Systems and Applications

PB - Association for Computing Machinery, Inc

ER -