The stapl Skeleton Framework

Mani Zandifar; Nathan Thomas; Nancy M. Amato; Lawrence Rauchwerger

doi:10.1007/978-3-319-17473-0_12

The stapl Skeleton Framework

Mani Zandifar, Nathan Thomas, Nancy M. Amato, Lawrence Rauchwerger

Research output: Chapter in Book/Report/Conference proceeding › Chapter

4 Scopus citations

Abstract

© Springer International Publishing Switzerland 2015. This paper describes the stapl Skeleton Framework, a highlevel skeletal approach for parallel programming. This framework abstracts the underlying details of data distribution and parallelism from programmers and enables them to express parallel programs as a composition of existing elementary skeletons such as map, map-reduce, scan, zip, butterfly, allreduce, alltoall and user-defined custom skeletons. Skeletons in this framework are defined as parametric data flow graphs, and their compositions are defined in terms of data flow graph compositions. Defining the composition in this manner allows dependencies between skeletons to be defined in terms of point-to-point dependencies, avoiding unnecessary global synchronizations. To show the ease of composability and expressivity, we implemented the NAS Integer Sort (IS) and Embarrassingly Parallel (EP) benchmarks using skeletons and demonstrate comparable performance to the hand-optimized reference implementations. To demonstrate scalable performance, we show a transformation which enables applications written in terms of skeletons to run on more than 100,000 cores.

Original language	English (US)
Title of host publication	Languages and Compilers for Parallel Computing
Publisher	Springer Nature
Pages	176-190
Number of pages	15
ISBN (Print)	9783319174723
DOIs	https://doi.org/10.1007/978-3-319-17473-0_12
State	Published - May 1 2015
Externally published	Yes

Bibliographical note

KAUST Repository Item: Exported on 2020-10-01
Acknowledged KAUST grant number(s): KUS-C1-016-04
Acknowledgements: This research supported in part by NSF awards CNS-0551685, CCF-0833199, CCF-0830753, IIS-0916053, IIS-0917266, EFRI-1240483, RI-1217991, by NIH NCI R25 CA090301-11, by DOE awards DE-AC02-06CH11357, DE-NA0002376, B575363, by Samsung, Chevron, IBM, Intel, Oracle/Sun and by Award KUS-C1-016-04, made by King Abdullah University of Science and Technology (KAUST). This research used resources of the National Energy Research Scientific Computing Center, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.
This publication acknowledges KAUST support, but has no KAUST affiliated authors.

Access to Document

10.1007/978-3-319-17473-0_12

Cite this

@inbook{a3504034917a4b3dba407652e82cc722,

title = "The stapl Skeleton Framework",

abstract = "{\textcopyright} Springer International Publishing Switzerland 2015. This paper describes the stapl Skeleton Framework, a highlevel skeletal approach for parallel programming. This framework abstracts the underlying details of data distribution and parallelism from programmers and enables them to express parallel programs as a composition of existing elementary skeletons such as map, map-reduce, scan, zip, butterfly, allreduce, alltoall and user-defined custom skeletons. Skeletons in this framework are defined as parametric data flow graphs, and their compositions are defined in terms of data flow graph compositions. Defining the composition in this manner allows dependencies between skeletons to be defined in terms of point-to-point dependencies, avoiding unnecessary global synchronizations. To show the ease of composability and expressivity, we implemented the NAS Integer Sort (IS) and Embarrassingly Parallel (EP) benchmarks using skeletons and demonstrate comparable performance to the hand-optimized reference implementations. To demonstrate scalable performance, we show a transformation which enables applications written in terms of skeletons to run on more than 100,000 cores.",

author = "Mani Zandifar and Nathan Thomas and Amato, {Nancy M.} and Lawrence Rauchwerger",

note = "KAUST Repository Item: Exported on 2020-10-01 Acknowledged KAUST grant number(s): KUS-C1-016-04 Acknowledgements: This research supported in part by NSF awards CNS-0551685, CCF-0833199, CCF-0830753, IIS-0916053, IIS-0917266, EFRI-1240483, RI-1217991, by NIH NCI R25 CA090301-11, by DOE awards DE-AC02-06CH11357, DE-NA0002376, B575363, by Samsung, Chevron, IBM, Intel, Oracle/Sun and by Award KUS-C1-016-04, made by King Abdullah University of Science and Technology (KAUST). This research used resources of the National Energy Research Scientific Computing Center, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231. This publication acknowledges KAUST support, but has no KAUST affiliated authors.",

year = "2015",

month = may,

day = "1",

doi = "10.1007/978-3-319-17473-0_12",

language = "English (US)",

isbn = "9783319174723",

pages = "176--190",

booktitle = "Languages and Compilers for Parallel Computing",

publisher = "Springer Nature",

}

TY - CHAP

T1 - The stapl Skeleton Framework

AU - Zandifar, Mani

AU - Thomas, Nathan

AU - Amato, Nancy M.

AU - Rauchwerger, Lawrence

N1 - KAUST Repository Item: Exported on 2020-10-01 Acknowledged KAUST grant number(s): KUS-C1-016-04 Acknowledgements: This research supported in part by NSF awards CNS-0551685, CCF-0833199, CCF-0830753, IIS-0916053, IIS-0917266, EFRI-1240483, RI-1217991, by NIH NCI R25 CA090301-11, by DOE awards DE-AC02-06CH11357, DE-NA0002376, B575363, by Samsung, Chevron, IBM, Intel, Oracle/Sun and by Award KUS-C1-016-04, made by King Abdullah University of Science and Technology (KAUST). This research used resources of the National Energy Research Scientific Computing Center, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231. This publication acknowledges KAUST support, but has no KAUST affiliated authors.

PY - 2015/5/1

Y1 - 2015/5/1

N2 - © Springer International Publishing Switzerland 2015. This paper describes the stapl Skeleton Framework, a highlevel skeletal approach for parallel programming. This framework abstracts the underlying details of data distribution and parallelism from programmers and enables them to express parallel programs as a composition of existing elementary skeletons such as map, map-reduce, scan, zip, butterfly, allreduce, alltoall and user-defined custom skeletons. Skeletons in this framework are defined as parametric data flow graphs, and their compositions are defined in terms of data flow graph compositions. Defining the composition in this manner allows dependencies between skeletons to be defined in terms of point-to-point dependencies, avoiding unnecessary global synchronizations. To show the ease of composability and expressivity, we implemented the NAS Integer Sort (IS) and Embarrassingly Parallel (EP) benchmarks using skeletons and demonstrate comparable performance to the hand-optimized reference implementations. To demonstrate scalable performance, we show a transformation which enables applications written in terms of skeletons to run on more than 100,000 cores.

AB - © Springer International Publishing Switzerland 2015. This paper describes the stapl Skeleton Framework, a highlevel skeletal approach for parallel programming. This framework abstracts the underlying details of data distribution and parallelism from programmers and enables them to express parallel programs as a composition of existing elementary skeletons such as map, map-reduce, scan, zip, butterfly, allreduce, alltoall and user-defined custom skeletons. Skeletons in this framework are defined as parametric data flow graphs, and their compositions are defined in terms of data flow graph compositions. Defining the composition in this manner allows dependencies between skeletons to be defined in terms of point-to-point dependencies, avoiding unnecessary global synchronizations. To show the ease of composability and expressivity, we implemented the NAS Integer Sort (IS) and Embarrassingly Parallel (EP) benchmarks using skeletons and demonstrate comparable performance to the hand-optimized reference implementations. To demonstrate scalable performance, we show a transformation which enables applications written in terms of skeletons to run on more than 100,000 cores.

UR - http://hdl.handle.net/10754/599966

UR - http://link.springer.com/10.1007/978-3-319-17473-0_12

UR - http://www.scopus.com/inward/record.url?scp=84937509918&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-17473-0_12

DO - 10.1007/978-3-319-17473-0_12

M3 - Chapter

SN - 9783319174723

SP - 176

EP - 190

BT - Languages and Compilers for Parallel Computing

PB - Springer Nature

ER -

The stapl Skeleton Framework

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this