A Fuzzy Neural Network Based Dynamic Data Allocation Model on Heterogeneous Multi-GPUs for Large-Scale Computations

Chaolong Zhang, Yuanping Xu, Zhijie Xu, Jia He, Jing Wang, Jianhua Adu

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

The parallel computation capabilities of modern graphics processing units (GPUs) have attracted increasing attention from researchers and engineers who have been conducting high computational throughput studies. However, current single GPU based engineering solutions are often struggling to fulfill their real-time requirements. Thus, the multi-GPU-based approach has become a popular and cost-effective choice for tackling the demands. In those cases, the computational load balancing over multiple GPU “nodes” is often the key and bottleneck that affect the quality and performance of the real-time system. The existing load balancing approaches are mainly based on the assumption that all GPU nodes in the same computer framework are of equal computational performance, which is often not the case due to cluster design and other legacy issues. This paper presents a novel dynamic load balancing (DLB) model for rapid data division and allocation on heterogeneous GPU nodes based on an innovative fuzzy neural network (FNN). In this research, a 5-state parameter feedback mechanism defining the overall cluster and node performance is proposed. The corresponding FNN-based DLB model will be capable of monitoring and predicting individual node performance under different workload scenarios. A real-time adaptive scheduler has been devised to reorganize the data inputs to each node when necessary to maintain their runtime computational performance. The devised model has been implemented on two dimensional (2D) discrete wavelet transform (DWT) applications for evaluation. Experiment results show that this DLB model enables a high computational throughput while ensuring real-time and precision requirements from complex computational tasks.

LanguageEnglish
Pages181-193
Number of pages13
JournalInternational Journal of Automation and Computing
Volume15
Issue number2
Early online date12 Mar 2018
DOIs
Publication statusPublished - Apr 2018

Fingerprint

Data Allocation
Fuzzy neural networks
Fuzzy Neural Network
Graphics Processing Unit
Resource allocation
Dynamic Load Balancing
Dynamic loads
Vertex of a graph
Real-time
Load Balancing
Throughput
Model
Discrete wavelet transforms
Requirements
Parallel Computation
Real time systems
Scheduler
Wavelet Transform
Workload
Graphics processing unit

Cite this

@article{5d5ab3347c944c2e899481e51e40867e,
title = "A Fuzzy Neural Network Based Dynamic Data Allocation Model on Heterogeneous Multi-GPUs for Large-Scale Computations",
abstract = "The parallel computation capabilities of modern graphics processing units (GPUs) have attracted increasing attention from researchers and engineers who have been conducting high computational throughput studies. However, current single GPU based engineering solutions are often struggling to fulfill their real-time requirements. Thus, the multi-GPU-based approach has become a popular and cost-effective choice for tackling the demands. In those cases, the computational load balancing over multiple GPU “nodes” is often the key and bottleneck that affect the quality and performance of the real-time system. The existing load balancing approaches are mainly based on the assumption that all GPU nodes in the same computer framework are of equal computational performance, which is often not the case due to cluster design and other legacy issues. This paper presents a novel dynamic load balancing (DLB) model for rapid data division and allocation on heterogeneous GPU nodes based on an innovative fuzzy neural network (FNN). In this research, a 5-state parameter feedback mechanism defining the overall cluster and node performance is proposed. The corresponding FNN-based DLB model will be capable of monitoring and predicting individual node performance under different workload scenarios. A real-time adaptive scheduler has been devised to reorganize the data inputs to each node when necessary to maintain their runtime computational performance. The devised model has been implemented on two dimensional (2D) discrete wavelet transform (DWT) applications for evaluation. Experiment results show that this DLB model enables a high computational throughput while ensuring real-time and precision requirements from complex computational tasks.",
keywords = "Heterogeneous GPU Cluster, Dynamic Load Balancing, Fuzzy Neural Network, Adaptive Scheduler, Discrete Wavelet Transform",
author = "Chaolong Zhang and Yuanping Xu and Zhijie Xu and Jia He and Jing Wang and Jianhua Adu",
year = "2018",
month = "4",
doi = "10.1007/s11633-018-1120-4",
language = "English",
volume = "15",
pages = "181--193",
journal = "International Journal of Automation and Computing",
issn = "1476-8186",
publisher = "Chinese Academy of Sciences",
number = "2",

}

A Fuzzy Neural Network Based Dynamic Data Allocation Model on Heterogeneous Multi-GPUs for Large-Scale Computations. / Zhang, Chaolong; Xu, Yuanping; Xu, Zhijie; He, Jia; Wang, Jing; Adu, Jianhua.

In: International Journal of Automation and Computing, Vol. 15, No. 2, 04.2018, p. 181-193.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A Fuzzy Neural Network Based Dynamic Data Allocation Model on Heterogeneous Multi-GPUs for Large-Scale Computations

AU - Zhang, Chaolong

AU - Xu, Yuanping

AU - Xu, Zhijie

AU - He, Jia

AU - Wang, Jing

AU - Adu, Jianhua

PY - 2018/4

Y1 - 2018/4

N2 - The parallel computation capabilities of modern graphics processing units (GPUs) have attracted increasing attention from researchers and engineers who have been conducting high computational throughput studies. However, current single GPU based engineering solutions are often struggling to fulfill their real-time requirements. Thus, the multi-GPU-based approach has become a popular and cost-effective choice for tackling the demands. In those cases, the computational load balancing over multiple GPU “nodes” is often the key and bottleneck that affect the quality and performance of the real-time system. The existing load balancing approaches are mainly based on the assumption that all GPU nodes in the same computer framework are of equal computational performance, which is often not the case due to cluster design and other legacy issues. This paper presents a novel dynamic load balancing (DLB) model for rapid data division and allocation on heterogeneous GPU nodes based on an innovative fuzzy neural network (FNN). In this research, a 5-state parameter feedback mechanism defining the overall cluster and node performance is proposed. The corresponding FNN-based DLB model will be capable of monitoring and predicting individual node performance under different workload scenarios. A real-time adaptive scheduler has been devised to reorganize the data inputs to each node when necessary to maintain their runtime computational performance. The devised model has been implemented on two dimensional (2D) discrete wavelet transform (DWT) applications for evaluation. Experiment results show that this DLB model enables a high computational throughput while ensuring real-time and precision requirements from complex computational tasks.

AB - The parallel computation capabilities of modern graphics processing units (GPUs) have attracted increasing attention from researchers and engineers who have been conducting high computational throughput studies. However, current single GPU based engineering solutions are often struggling to fulfill their real-time requirements. Thus, the multi-GPU-based approach has become a popular and cost-effective choice for tackling the demands. In those cases, the computational load balancing over multiple GPU “nodes” is often the key and bottleneck that affect the quality and performance of the real-time system. The existing load balancing approaches are mainly based on the assumption that all GPU nodes in the same computer framework are of equal computational performance, which is often not the case due to cluster design and other legacy issues. This paper presents a novel dynamic load balancing (DLB) model for rapid data division and allocation on heterogeneous GPU nodes based on an innovative fuzzy neural network (FNN). In this research, a 5-state parameter feedback mechanism defining the overall cluster and node performance is proposed. The corresponding FNN-based DLB model will be capable of monitoring and predicting individual node performance under different workload scenarios. A real-time adaptive scheduler has been devised to reorganize the data inputs to each node when necessary to maintain their runtime computational performance. The devised model has been implemented on two dimensional (2D) discrete wavelet transform (DWT) applications for evaluation. Experiment results show that this DLB model enables a high computational throughput while ensuring real-time and precision requirements from complex computational tasks.

KW - Heterogeneous GPU Cluster

KW - Dynamic Load Balancing

KW - Fuzzy Neural Network

KW - Adaptive Scheduler

KW - Discrete Wavelet Transform

UR - http://www.scopus.com/inward/record.url?scp=85043379198&partnerID=8YFLogxK

U2 - 10.1007/s11633-018-1120-4

DO - 10.1007/s11633-018-1120-4

M3 - Article

VL - 15

SP - 181

EP - 193

JO - International Journal of Automation and Computing

T2 - International Journal of Automation and Computing

JF - International Journal of Automation and Computing

SN - 1476-8186

IS - 2

ER -