MapReduce for the Cell B.E. Architecture
| Sorted by Date | Classified by Publication Type | Classified by Project |
Marc de Kruijf and Karthikeyan Sankaralingam. MapReduce for the Cell B.E. Architecture. Technical Report TR1625, Department of Computer Sciences, The University of Wisconsin-Madison, 2007.
Won 2nd Place in IBM Cell challenge 2007.
Download
Abstract
MapReduce is a simple and flexible parallel programming model proposedby Google for large scale data processing in a distributed computingenvironment [4]. In this paper, we present a design and implementationof MapReduce for the Cell architecture. This model provides a simplemachine abstraction to users, hiding parallelization and hardwareprimitives. Our runtime automatically manages parallelization,scheduling, partitioning and memory transfers. We study the basiccharacteristics of the model and evaluate our runtime's performance,scalability, and efficiency for micro-benchmarks and completeapplications.We show that the model is well suited for manyapplications that map well to the Cell architecture, and that theruntime sustains high performance on these applications. For otherapplications, we analyze runtime performance and describe whyperformance is less impressive. Overall, we find that the simplicityof the model and the efficiency of our MapReduce implementationmake itan attractive choice for the Cell platform specifically and moregenerally to distributed memory systems and software-exposed memories.
Additional Information
This paper has been submitted for publication in the IBM Journal of Research and Development. 2nd Place in IBM Cell challenge 2007. http://www.cs.wisc.edu/vertical/archive/cell-challenge-2007/22363.wss
BibTeX
@TECHREPORT{MapreduceCell2007,
AUTHOR = {Marc de Kruijf and Karthikeyan Sankaralingam},
TITLE = "{MapReduce for the Cell B.E. Architecture}",
abstract = {
MapReduce is a simple and flexible parallel programming model proposed
by Google for large scale data processing in a distributed computing
environment [4]. In this paper, we present a design and implementation
of MapReduce for the Cell architecture. This model provides a simple
machine abstraction to users, hiding parallelization and hardware
primitives. Our runtime automatically manages parallelization,
scheduling, partitioning and memory transfers. We study the basic
characteristics of the model and evaluate our runtime's performance,
scalability, and efficiency for micro-benchmarks and complete
applications.We show that the model is well suited for many
applications that map well to the Cell architecture, and that the
runtime sustains high performance on these applications. For other
applications, we analyze runtime performance and describe why
performance is less impressive. Overall, we find that the simplicity
of the model and the efficiency of our MapReduce implementationmake it
an attractive choice for the Cell platform specifically and more
generally to distributed memory systems and software-exposed memories.
},
INSTITUTION = {Department of Computer Sciences, The University of Wisconsin-Madison},
SCHOOL = {The University of Wisconsin-Madison},
ADDRESS = {Madison, WI},
YEAR = 2007,
NUMBER = {TR1625}
bib_dl = {http://www.cs.wisc.edu/techreports/viewreport.php?report=1625},
wwwnote = {Won <b>2nd Place in IBM Cell challenge 2007.</b>},
bib_dl_pdf = {http://www.cs.wisc.edu/techreports/2007/TR1625.pdf},
bib_pubtype = {Other,Award Paper},
bib_rescat = {proj-opinion},
bib_extra_info = {This paper has been submitted for publication in the IBM Journal of Research and Development. 2nd Place in IBM Cell challenge 2007. http://www.cs.wisc.edu/vertical/archive/cell-challenge-2007/22363.wss }
}
Generated by bib.pl (written by Patrick Riley ) on Sun Sep 26, 2021 16:14:28 time=1207019082