Time Performance Analysis of Multi-CPU and Multi-GPU in Big Data Clustering Computation

被引：0

作者：

Adiyoso, Widiarto ^{[1
]}

Krisnadhi, Adila ^{[1
]}

Wibisono, Ari ^{[1
]}

Purbarani, Sumarsih Condroayu ^{[1
]}

Saraswati, Anindhita Dwi ^{[1
]}

Putri, Annissa Fildzah Rafi ^{[1
]}

Saladdin, Ibad Rahadian ^{[1
]}

Anwar, S. Reyneta Carissa ^{[1
]}

机构：

[1] Univ Indonesia, Fac Comp Sci, Depok, Indonesia

来源：

2018 INTERNATIONAL WORKSHOP ON BIG DATA AND INFORMATION SECURITY (IWBIS) | 2018年

关键词：

Spark; TensorFlow; Big Data; Cluster; K-Means; OPTIMIZATION;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Big data is a hot topic that is regularly discussed in the computer science field for the past year. Big data provides numerous benefits for the development of technologies, such as business intelligence and deep learning. Processing big data requires specialized tools and environment, ranging from a commodity-clustered workstation to high performance computing server, especially in big data clustering where unsupervised learning takes place. In this paper, we conduct time analysis of commodity-clustered workstation equipped with Spark as a baseline for multi-CPU big data clustering and TensorFlow installed in a high-performance computing workstation as a baseline for multi-GPU big data clustering. Based on the analysis, it shows that TensorFlow performs have around 5 to 12 times faster computation time than Spark.

引用

页码：113 / 116

页数：4

共 50 条

[1] Financial applications on multi-CPU and multi-GPU architectures
Emilio Castillo
Cristóbal Camarero
Ana Borrego
Jose Luis Bosque
[J]. The Journal of Supercomputing, 2015, 71 : 729 - 739
[2] Financial applications on multi-CPU and multi-GPU architectures
Castillo, Emilio
Camarero, Cristobal
Borrego, Ana
Luis Bosque, Jose
[J]. JOURNAL OF SUPERCOMPUTING, 2015, 71 (02): : 729 - 739
[3] Financial applications on multi-CPU and multi-GPU architectures
Department of Computer Science and Electronics, Universidad de Cantabria, Santander, Spain
不详
[J]. J Supercomput, 2 (729-739):
[4] Design and analysis of scheduling strategies for multi-CPU and multi-GPU architectures
Lima, Joao V. F.
Gautier, Thierry
Danjean, Vincent
Raffin, Bruno
Maillard, Nicolas
[J]. PARALLEL COMPUTING, 2015, 44 : 37 - 52
[5] Multi-GPU and Multi-CPU Parallelization for Interactive Physics Simulations
Hermann, Everton
Raffin, Bruno
Faure, Francois
Gautier, Thierry
Allard, Jeremie
[J]. EURO-PAR 2010 - PARALLEL PROCESSING, PART II, 2010, 6272 : 235 - 246
[6] Multi-CPU/Multi-GPU Based Framework for Multimedia Processing
Mahmoudi, Sidi Ahmed
Manneback, Pierre
[J]. COMPUTER SCIENCE AND ITS APPLICATIONS, CIIA 2015, 2015, 456 : 54 - 65
[7] HPSM: A Programming Framework for Multi-CPU and Multi-GPU Systems
Lima, Joao V. F.
Di Domenico, Daniel
[J]. 2017 INTERNATIONAL SYMPOSIUM ON COMPUTER ARCHITECTURE AND HIGH PERFORMANCE COMPUTING WORKSHOPS (SBAC-PADW), 2017, : 31 - 36
[8] Strategies for maximizing utilization on multi-CPU and multi-GPU heterogeneous architectures
Angeles Navarro
Antonio Vilches
Francisco Corbera
Rafael Asenjo
[J]. The Journal of Supercomputing, 2014, 70 : 756 - 771
[9] Multi-GPU and multi-CPU accelerated FDTD scheme for vibroacoustic applications
Frances, J.
Otero, B.
Bleda, S.
Gallego, S.
Neipp, C.
Marquez, A.
Belendez, A.
[J]. COMPUTER PHYSICS COMMUNICATIONS, 2015, 191 : 43 - 51
[10] Strategies for maximizing utilization on multi-CPU and multi-GPU heterogeneous architectures
Navarro, Angeles
Vilches, Antonio
Corbera, Francisco
Asenjo, Rafael
[J]. JOURNAL OF SUPERCOMPUTING, 2014, 70 (02): : 756 - 771

← 1 2 3 4 5 →