A hybrid shared memory execution model for a data parallel language with I/O

Clemens Grelck; Steffen Kuthe; Sven-Bodo Scholz

doi:10.1142/S012962640800320X

A hybrid shared memory execution model for a data parallel language with I/O

Clemens Grelck, Steffen Kuthe, Sven-Bodo Scholz

Research output: Contribution to journal › Article › peer-review

2 Citations (Scopus)

Abstract

We propose a novel execution model for the implicitly parallel execution of data parallel programs in the presence of general I/O operations. This model is called hybrid because it combines the advantages of the standard execution models fork/join and SPMD. Based on program analysis the hybrid model adapts itself to one or the other on the granularity of individual instructions. We outline compilation techniques that systematically derive the organization of parallel code from data flow characteristics aiming at the reduction of execution mode switches in general and synchronization/communication requirements in particular. Experiments based on a prototype implementation show the effectiveness of the hybrid execution model for reducing parallel overhead.

Original language	English
Pages (from-to)	23-37
Journal	Parallel Processing Letters
Volume	18
Issue number	1
DOIs	https://doi.org/10.1142/S012962640800320X
Publication status	Published - 2008

Keywords

multithreaded execution
execution models
implicit parallelisation
single assignment C
declarative array programming
data parallel programming

Access to Document

10.1142/S012962640800320X

Cite this

@article{7ebc99a293aa40cba8ead1e501af1153,

title = "A hybrid shared memory execution model for a data parallel language with I/O",

abstract = "We propose a novel execution model for the implicitly parallel execution of data parallel programs in the presence of general I/O operations. This model is called hybrid because it combines the advantages of the standard execution models fork/join and SPMD. Based on program analysis the hybrid model adapts itself to one or the other on the granularity of individual instructions. We outline compilation techniques that systematically derive the organization of parallel code from data flow characteristics aiming at the reduction of execution mode switches in general and synchronization/communication requirements in particular. Experiments based on a prototype implementation show the effectiveness of the hybrid execution model for reducing parallel overhead.",

keywords = "multithreaded execution, execution models, implicit parallelisation, single assignment C, declarative array programming, data parallel programming",

author = "Clemens Grelck and Steffen Kuthe and Sven-Bodo Scholz",

note = "Original article can be found at: http://www.worldscinet.com/ Copyright World Scientific Publishing Company [Full text of this article is not available in the UHRA]",

year = "2008",

doi = "10.1142/S012962640800320X",

language = "English",

volume = "18",

pages = "23--37",

journal = "Parallel Processing Letters",

issn = "0129-6264",

publisher = "World Scientific Publishing Co. Pte Ltd",

number = "1",

}

TY - JOUR

T1 - A hybrid shared memory execution model for a data parallel language with I/O

AU - Grelck, Clemens

AU - Kuthe, Steffen

AU - Scholz, Sven-Bodo

N1 - Original article can be found at: http://www.worldscinet.com/ Copyright World Scientific Publishing Company [Full text of this article is not available in the UHRA]

PY - 2008

Y1 - 2008

N2 - We propose a novel execution model for the implicitly parallel execution of data parallel programs in the presence of general I/O operations. This model is called hybrid because it combines the advantages of the standard execution models fork/join and SPMD. Based on program analysis the hybrid model adapts itself to one or the other on the granularity of individual instructions. We outline compilation techniques that systematically derive the organization of parallel code from data flow characteristics aiming at the reduction of execution mode switches in general and synchronization/communication requirements in particular. Experiments based on a prototype implementation show the effectiveness of the hybrid execution model for reducing parallel overhead.

AB - We propose a novel execution model for the implicitly parallel execution of data parallel programs in the presence of general I/O operations. This model is called hybrid because it combines the advantages of the standard execution models fork/join and SPMD. Based on program analysis the hybrid model adapts itself to one or the other on the granularity of individual instructions. We outline compilation techniques that systematically derive the organization of parallel code from data flow characteristics aiming at the reduction of execution mode switches in general and synchronization/communication requirements in particular. Experiments based on a prototype implementation show the effectiveness of the hybrid execution model for reducing parallel overhead.

KW - multithreaded execution

KW - execution models

KW - implicit parallelisation

KW - single assignment C

KW - declarative array programming

KW - data parallel programming

U2 - 10.1142/S012962640800320X

DO - 10.1142/S012962640800320X

M3 - Article

SN - 0129-6264

VL - 18

SP - 23

EP - 37

JO - Parallel Processing Letters

JF - Parallel Processing Letters

IS - 1

ER -

A hybrid shared memory execution model for a data parallel language with I/O

Abstract

Keywords

Access to Document

Fingerprint

Cite this