Dynamic Load Balancing on Multi-GPUs System for Big Data Processing

Chaolong Zhang, Yuanping Xu, Jiliu Zhou, Zhijie Xu, Li Lu, Jun Lu

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

1 Citation (Scopus)

Abstract

The powerful parallel computing capability of modern GPU (Graphics Processing Unit) processors has attracted increasing attentions of researchers and engineers who had conducted a large number of GPU-based acceleration research projects. However, current single GPU based solutions are still incapable of fulfilling the real-time computational requirements from the latest big data applications. Thus, the multi-GPU solution has become a trend for many real-time application attempts. In those cases, the computational load balancing over the multiple GPU nodes is often the key bottleneck that needs to be further studied to ensure the best possible performance. The existing load balancing approaches are mainly based on the assumption that all GPUs in the same system provide equal computational performance, and had fallen short to address the situations from heterogeneous multiGPU systems. This paper presents a novel dynamic load balancing model for heterogeneous multi-GPU systems based on the fuzzy neural network (FNN) framework. The devised model has been implemented and demonstrated in a case study for improving the computational performance of a two dimensional (2D) discrete wavelet transform (DWT). Experiment results show that this dynamic load balancing model has enabled a high computational throughput that can satisfy the real-time and accuracy requirements from many big data processing applications
Original languageEnglish
Title of host publicationProceedings of the 23rd International Conference on Automation & Computing, (University of Huddersfield, 7-8 September 2017)
PublisherInstitute of Electrical and Electronics Engineers Inc.
Number of pages6
ISBN (Electronic)9780701702601
ISBN (Print)9781509050406
DOIs
Publication statusPublished - 26 Oct 2017
Event23rd International Conference on Automation and Computing: Addressing Global Challenges through Automation and Computing - University of Huddersfield, Huddersfield, United Kingdom
Duration: 7 Sep 20178 Sep 2017
Conference number: 23
https://www.ieee.org/conferences_events/conferences/conferencedetails/index.html?Conf_ID=41042 (Link to Conference Website)

Conference

Conference23rd International Conference on Automation and Computing
Abbreviated titleICAC 2017
CountryUnited Kingdom
CityHuddersfield
Period7/09/178/09/17
OtherThe scope of the conference covers a broad spectrum of areas with multi-disciplinary interests in the fields of automation, control engineering, computing and information systems, ranging from fundamental research to real-world applications.
Internet address

Fingerprint

Dynamic loads
Resource allocation
Fuzzy neural networks
Discrete wavelet transforms
Parallel processing systems
Big data
Graphics processing unit
Throughput
Engineers
Experiments

Cite this

Zhang, C., Xu, Y., Zhou, J., Xu, Z., Lu, L., & Lu, J. (2017). Dynamic Load Balancing on Multi-GPUs System for Big Data Processing. In Proceedings of the 23rd International Conference on Automation & Computing, (University of Huddersfield, 7-8 September 2017) Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.23919/IConAC.2017.8082085
Zhang, Chaolong ; Xu, Yuanping ; Zhou, Jiliu ; Xu, Zhijie ; Lu, Li ; Lu, Jun. / Dynamic Load Balancing on Multi-GPUs System for Big Data Processing. Proceedings of the 23rd International Conference on Automation & Computing, (University of Huddersfield, 7-8 September 2017). Institute of Electrical and Electronics Engineers Inc., 2017.
@inproceedings{1d7610759bb14d8584c290a7d152ad7c,
title = "Dynamic Load Balancing on Multi-GPUs System for Big Data Processing",
abstract = "The powerful parallel computing capability of modern GPU (Graphics Processing Unit) processors has attracted increasing attentions of researchers and engineers who had conducted a large number of GPU-based acceleration research projects. However, current single GPU based solutions are still incapable of fulfilling the real-time computational requirements from the latest big data applications. Thus, the multi-GPU solution has become a trend for many real-time application attempts. In those cases, the computational load balancing over the multiple GPU nodes is often the key bottleneck that needs to be further studied to ensure the best possible performance. The existing load balancing approaches are mainly based on the assumption that all GPUs in the same system provide equal computational performance, and had fallen short to address the situations from heterogeneous multiGPU systems. This paper presents a novel dynamic load balancing model for heterogeneous multi-GPU systems based on the fuzzy neural network (FNN) framework. The devised model has been implemented and demonstrated in a case study for improving the computational performance of a two dimensional (2D) discrete wavelet transform (DWT). Experiment results show that this dynamic load balancing model has enabled a high computational throughput that can satisfy the real-time and accuracy requirements from many big data processing applications",
keywords = "Multi-GPU, Load balancing, Fuzzy neural network, DWT",
author = "Chaolong Zhang and Yuanping Xu and Jiliu Zhou and Zhijie Xu and Li Lu and Jun Lu",
year = "2017",
month = "10",
day = "26",
doi = "10.23919/IConAC.2017.8082085",
language = "English",
isbn = "9781509050406",
booktitle = "Proceedings of the 23rd International Conference on Automation & Computing, (University of Huddersfield, 7-8 September 2017)",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

Zhang, C, Xu, Y, Zhou, J, Xu, Z, Lu, L & Lu, J 2017, Dynamic Load Balancing on Multi-GPUs System for Big Data Processing. in Proceedings of the 23rd International Conference on Automation & Computing, (University of Huddersfield, 7-8 September 2017). Institute of Electrical and Electronics Engineers Inc., 23rd International Conference on Automation and Computing, Huddersfield, United Kingdom, 7/09/17. https://doi.org/10.23919/IConAC.2017.8082085

Dynamic Load Balancing on Multi-GPUs System for Big Data Processing. / Zhang, Chaolong; Xu, Yuanping; Zhou, Jiliu; Xu, Zhijie; Lu, Li; Lu, Jun.

Proceedings of the 23rd International Conference on Automation & Computing, (University of Huddersfield, 7-8 September 2017). Institute of Electrical and Electronics Engineers Inc., 2017.

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

TY - GEN

T1 - Dynamic Load Balancing on Multi-GPUs System for Big Data Processing

AU - Zhang, Chaolong

AU - Xu, Yuanping

AU - Zhou, Jiliu

AU - Xu, Zhijie

AU - Lu, Li

AU - Lu, Jun

PY - 2017/10/26

Y1 - 2017/10/26

N2 - The powerful parallel computing capability of modern GPU (Graphics Processing Unit) processors has attracted increasing attentions of researchers and engineers who had conducted a large number of GPU-based acceleration research projects. However, current single GPU based solutions are still incapable of fulfilling the real-time computational requirements from the latest big data applications. Thus, the multi-GPU solution has become a trend for many real-time application attempts. In those cases, the computational load balancing over the multiple GPU nodes is often the key bottleneck that needs to be further studied to ensure the best possible performance. The existing load balancing approaches are mainly based on the assumption that all GPUs in the same system provide equal computational performance, and had fallen short to address the situations from heterogeneous multiGPU systems. This paper presents a novel dynamic load balancing model for heterogeneous multi-GPU systems based on the fuzzy neural network (FNN) framework. The devised model has been implemented and demonstrated in a case study for improving the computational performance of a two dimensional (2D) discrete wavelet transform (DWT). Experiment results show that this dynamic load balancing model has enabled a high computational throughput that can satisfy the real-time and accuracy requirements from many big data processing applications

AB - The powerful parallel computing capability of modern GPU (Graphics Processing Unit) processors has attracted increasing attentions of researchers and engineers who had conducted a large number of GPU-based acceleration research projects. However, current single GPU based solutions are still incapable of fulfilling the real-time computational requirements from the latest big data applications. Thus, the multi-GPU solution has become a trend for many real-time application attempts. In those cases, the computational load balancing over the multiple GPU nodes is often the key bottleneck that needs to be further studied to ensure the best possible performance. The existing load balancing approaches are mainly based on the assumption that all GPUs in the same system provide equal computational performance, and had fallen short to address the situations from heterogeneous multiGPU systems. This paper presents a novel dynamic load balancing model for heterogeneous multi-GPU systems based on the fuzzy neural network (FNN) framework. The devised model has been implemented and demonstrated in a case study for improving the computational performance of a two dimensional (2D) discrete wavelet transform (DWT). Experiment results show that this dynamic load balancing model has enabled a high computational throughput that can satisfy the real-time and accuracy requirements from many big data processing applications

KW - Multi-GPU

KW - Load balancing

KW - Fuzzy neural network

KW - DWT

UR - http://ieeexplore.ieee.org/xpl/conhome.jsp?punumber=1800563

U2 - 10.23919/IConAC.2017.8082085

DO - 10.23919/IConAC.2017.8082085

M3 - Conference contribution

SN - 9781509050406

BT - Proceedings of the 23rd International Conference on Automation & Computing, (University of Huddersfield, 7-8 September 2017)

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Zhang C, Xu Y, Zhou J, Xu Z, Lu L, Lu J. Dynamic Load Balancing on Multi-GPUs System for Big Data Processing. In Proceedings of the 23rd International Conference on Automation & Computing, (University of Huddersfield, 7-8 September 2017). Institute of Electrical and Electronics Engineers Inc. 2017 https://doi.org/10.23919/IConAC.2017.8082085