DPM: Data Partitioning Method for pipelined MapReduce on GPU

Author

Myung Hyun Jo ; Won Woo Ro

Author_Institution

Sch. of Electr. & Electron. Eng., Yonsei Univ., Seoul, South Korea

fYear

2014

fDate

22-25 June 2014

Firstpage

1

Lastpage

3

Abstract

The MapReduce frameworks using a modern graphic processor (GPU) have improved the performance of data-intensive applications. While the prior researches have enhanced the parallelism of the MapReduce application on a GPU, archiving optimal distribution of big data on heterogeneous devices is still a challengeable issue. We therefore propose a method to evenly separate the computing cost under limited memory size. To solve this problem, we design and propose DPM, a Data Partitioning Method, using a GPU to smartly distribute workload of MapReduce. The proposed technique provides well-balanced processing cost for heterogeneous devices.

Keywords

Big Data; graphics processing units; parallel programming; pipeline processing; DPM; GPU; MapReduce application parallelism; data partitioning method; data-intensive applications; graphic processor; heterogeneous devices; memory size; optimal big data distribution; performance improvement; pipelined MapReduce; processing cost; workload distribution; Big data; Buffer storage; Computer architecture; Data models; Educational institutions; Graphics processing units; Optical wavelength conversion; GPU; MapReduce; big data; parallel; pipeline;

fLanguage

English

Publisher

ieee

Conference_Titel

Consumer Electronics (ISCE 2014), The 18th IEEE International Symposium on

Conference_Location

JeJu Island

Type

conf

DOI

10.1109/ISCE.2014.6884382

Filename

6884382