Cited 14 time in webofscience Cited 20 time in scopus

Grex: An efficient MapReduce framework for graphics processing units

Title
Grex: An efficient MapReduce framework for graphics processing units
Authors
Basaran, C[Basaran, Can]Kang, KD[Kang, Kyoung-Don]
DGIST Authors
Basaran, C[Basaran, Can]
Issue Date
2013-04
Citation
Journal of Parallel and Distributed Computing, 73(4), 522-533
Type
Article
Article Type
Article
Keywords
Computer GraphicsComputer Graphics EquipmentData ProcessingGPGPUGraphics Processing UnitGraphics Processing UnitsMap-ReduceMemory ArchitectureParallel Data ProcessingProgram ProcessorsShared MemorySIMTStorage Allocation (Computer)Thread Synchronization
ISSN
0743-7315
Abstract
In this paper, we present a new MapReduce framework, called Grex, designed to leverage general purpose graphics processing units (GPUs) for parallel data processing. Grex provides several new features. First, it supports a parallel split method to tokenize input data of variable sizes, such as words in e-books or URLs in web documents, in parallel using GPU threads. Second, Grex evenly distributes data to map/reduce tasks to avoid data partitioning skews. In addition, Grex provides a new memory management scheme to enhance the performance by exploiting the GPU memory hierarchy. Notably, all these capabilities are supported via careful system design without requiring any locks or atomic operations for thread synchronization. The experimental results show that our system is up to 12.4× and 4.1× faster than two state-of-the-art GPU-based MapReduce frameworks for the tested applications. © 2013 Elsevier Inc. All rights reserved.
URI
http://hdl.handle.net/20.500.11750/3244
DOI
10.1016/j.jpdc.2013.01.004
Publisher
Elsevier
Files:
There are no files associated with this item.
Collection:
Information and Communication EngineeringRTCPS(Real-Time Cyber-Physical Systems) Lab1. Journal Articles


qrcode mendeley

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE