Skip to main navigation Skip to search Skip to main content

Simulated Annealing for Timeliness and Energy aware Deep Learning Job Assignment

  • Korea Advanced Institute of Science and Technology

Research output: Contribution to conferenceConference paperpeer-review

Abstract

In this paper, we propose a simulated annealing method for timeliness and energy aware deep learning (DL) job assignment. In the proposed method, we make a decision of server position for DL training jobs in order to consider both the timeliness of deep neural network (DNN) model updating and the associated energy consumption. For timeliness management, we design three penalty functions; step, linear, and exponential functions, so as to practically penalize the service quality degradation of DL inference due to delay of DL training. For energy management, we formulate the operation cost optimization problem considering energy consumption for DL training. By using these management schemes, our method is able to find the trade-off of the timeliness and energy consumption of DL training. Especially, in order to overcome the non-smoothness of the defined problem, we design a simulated annealing (SA) based metaheuristic which finds approximated optimal solution for DL training job assignment. For the performance evaluation, we show the preliminary experimental results of DL training jobs with AlexNet, Inception-V3, and ResNet on NVIDIA GPU devices.

Original languageEnglish
Title of host publicationICTC 2019 - 10th International Conference on ICT Convergence
Subtitle of host publicationICT Convergence Leading the Autonomous Future
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages571-575
Number of pages5
ISBN (Electronic)9781728108926
DOIs
StatePublished - 2019.10
Event10th International Conference on Information and Communication Technology Convergence, ICTC 2019 - Jeju Island, Korea, Republic of
Duration: 2019.10.162019.10.18

Publication series

NameICTC 2019 - 10th International Conference on ICT Convergence: ICT Convergence Leading the Autonomous Future

Conference

Conference10th International Conference on Information and Communication Technology Convergence, ICTC 2019
Country/TerritoryKorea, Republic of
CityJeju Island
Period19.10.1619.10.18

Keywords

  • deep learning inference
  • deep learning training
  • deep neural network
  • energy consumption
  • power usage
  • timeliness

Fingerprint

Dive into the research topics of 'Simulated Annealing for Timeliness and Energy aware Deep Learning Job Assignment'. Together they form a unique fingerprint.

Cite this