Lazy parallel kronecker algebra-operations on heterogeneous multicores

Wasuwee Sodsong, Robert Mittermayr, Yoojin Park, Bernd Burgstaller, Johann Blieberger

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Kronecker algebra is a matrix calculus which allows the generation of thread interleavings from the source-code of a program. Thread interleavings have been shown effective for proving the absence of deadlocks. Because the number of interleavings grows exponentially in the number of threads, deadlock analysis is still a challenging problem. To make the computation of thread interleavings tractable, we propose a lazy, parallel evaluation method for Kronecker algebra. Our method incorporates the constraints induced by synchronization constructs. To reduce problem size, only interleavings legal under the locking behavior of a program are considered. We leverage the data-parallelism of Kronecker sum- and product-operations for multicores and GPUs. Proposed algebraic transformations further improve performance. For one synthetic and two real-world benchmarks, our GPU implementation is up to 5453 X faster than our multi-threaded version. Lazy evaluation significantly reduces memory consumption compared to both the sequential and the multicore versions of the SPIN model-checker.

Original languageEnglish
Title of host publicationEuro-Par 2017
Subtitle of host publicationParallel Processing - 23rd International Conference on Parallel and Distributed Computing, Proceedings
EditorsFrancisco F. Rivera, Tomas F. Pena, Jose C. Cabaleiro
PublisherSpringer Verlag
Pages538-552
Number of pages15
ISBN (Print)9783319642024
DOIs
Publication statusPublished - 2017 Jan 1
Event23rd International Conference on Parallel and Distributed Computing, Euro-Par 2017 - Santiago de Compostela, Spain
Duration: 2017 Aug 282017 Sep 1

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10417 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other23rd International Conference on Parallel and Distributed Computing, Euro-Par 2017
CountrySpain
CitySantiago de Compostela
Period17/8/2817/9/1

Fingerprint

Interleaving
Algebra
Thread
Deadlock
Synchronization
Data storage equipment
Data Parallelism
Parallel Methods
Spin Models
Locking
Evaluation Method
Leverage
Calculus
Graphics processing unit
Benchmark
Evaluation

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Sodsong, W., Mittermayr, R., Park, Y., Burgstaller, B., & Blieberger, J. (2017). Lazy parallel kronecker algebra-operations on heterogeneous multicores. In F. F. Rivera, T. F. Pena, & J. C. Cabaleiro (Eds.), Euro-Par 2017: Parallel Processing - 23rd International Conference on Parallel and Distributed Computing, Proceedings (pp. 538-552). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10417 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-64203-1_39
Sodsong, Wasuwee ; Mittermayr, Robert ; Park, Yoojin ; Burgstaller, Bernd ; Blieberger, Johann. / Lazy parallel kronecker algebra-operations on heterogeneous multicores. Euro-Par 2017: Parallel Processing - 23rd International Conference on Parallel and Distributed Computing, Proceedings. editor / Francisco F. Rivera ; Tomas F. Pena ; Jose C. Cabaleiro. Springer Verlag, 2017. pp. 538-552 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{9d4b76400eeb4398ab8af14b6b00d09b,
title = "Lazy parallel kronecker algebra-operations on heterogeneous multicores",
abstract = "Kronecker algebra is a matrix calculus which allows the generation of thread interleavings from the source-code of a program. Thread interleavings have been shown effective for proving the absence of deadlocks. Because the number of interleavings grows exponentially in the number of threads, deadlock analysis is still a challenging problem. To make the computation of thread interleavings tractable, we propose a lazy, parallel evaluation method for Kronecker algebra. Our method incorporates the constraints induced by synchronization constructs. To reduce problem size, only interleavings legal under the locking behavior of a program are considered. We leverage the data-parallelism of Kronecker sum- and product-operations for multicores and GPUs. Proposed algebraic transformations further improve performance. For one synthetic and two real-world benchmarks, our GPU implementation is up to 5453 X faster than our multi-threaded version. Lazy evaluation significantly reduces memory consumption compared to both the sequential and the multicore versions of the SPIN model-checker.",
author = "Wasuwee Sodsong and Robert Mittermayr and Yoojin Park and Bernd Burgstaller and Johann Blieberger",
year = "2017",
month = "1",
day = "1",
doi = "10.1007/978-3-319-64203-1_39",
language = "English",
isbn = "9783319642024",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "538--552",
editor = "Rivera, {Francisco F.} and Pena, {Tomas F.} and Cabaleiro, {Jose C.}",
booktitle = "Euro-Par 2017",
address = "Germany",

}

Sodsong, W, Mittermayr, R, Park, Y, Burgstaller, B & Blieberger, J 2017, Lazy parallel kronecker algebra-operations on heterogeneous multicores. in FF Rivera, TF Pena & JC Cabaleiro (eds), Euro-Par 2017: Parallel Processing - 23rd International Conference on Parallel and Distributed Computing, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10417 LNCS, Springer Verlag, pp. 538-552, 23rd International Conference on Parallel and Distributed Computing, Euro-Par 2017, Santiago de Compostela, Spain, 17/8/28. https://doi.org/10.1007/978-3-319-64203-1_39

Lazy parallel kronecker algebra-operations on heterogeneous multicores. / Sodsong, Wasuwee; Mittermayr, Robert; Park, Yoojin; Burgstaller, Bernd; Blieberger, Johann.

Euro-Par 2017: Parallel Processing - 23rd International Conference on Parallel and Distributed Computing, Proceedings. ed. / Francisco F. Rivera; Tomas F. Pena; Jose C. Cabaleiro. Springer Verlag, 2017. p. 538-552 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10417 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Lazy parallel kronecker algebra-operations on heterogeneous multicores

AU - Sodsong, Wasuwee

AU - Mittermayr, Robert

AU - Park, Yoojin

AU - Burgstaller, Bernd

AU - Blieberger, Johann

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Kronecker algebra is a matrix calculus which allows the generation of thread interleavings from the source-code of a program. Thread interleavings have been shown effective for proving the absence of deadlocks. Because the number of interleavings grows exponentially in the number of threads, deadlock analysis is still a challenging problem. To make the computation of thread interleavings tractable, we propose a lazy, parallel evaluation method for Kronecker algebra. Our method incorporates the constraints induced by synchronization constructs. To reduce problem size, only interleavings legal under the locking behavior of a program are considered. We leverage the data-parallelism of Kronecker sum- and product-operations for multicores and GPUs. Proposed algebraic transformations further improve performance. For one synthetic and two real-world benchmarks, our GPU implementation is up to 5453 X faster than our multi-threaded version. Lazy evaluation significantly reduces memory consumption compared to both the sequential and the multicore versions of the SPIN model-checker.

AB - Kronecker algebra is a matrix calculus which allows the generation of thread interleavings from the source-code of a program. Thread interleavings have been shown effective for proving the absence of deadlocks. Because the number of interleavings grows exponentially in the number of threads, deadlock analysis is still a challenging problem. To make the computation of thread interleavings tractable, we propose a lazy, parallel evaluation method for Kronecker algebra. Our method incorporates the constraints induced by synchronization constructs. To reduce problem size, only interleavings legal under the locking behavior of a program are considered. We leverage the data-parallelism of Kronecker sum- and product-operations for multicores and GPUs. Proposed algebraic transformations further improve performance. For one synthetic and two real-world benchmarks, our GPU implementation is up to 5453 X faster than our multi-threaded version. Lazy evaluation significantly reduces memory consumption compared to both the sequential and the multicore versions of the SPIN model-checker.

UR - http://www.scopus.com/inward/record.url?scp=85028730066&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85028730066&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-64203-1_39

DO - 10.1007/978-3-319-64203-1_39

M3 - Conference contribution

AN - SCOPUS:85028730066

SN - 9783319642024

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 538

EP - 552

BT - Euro-Par 2017

A2 - Rivera, Francisco F.

A2 - Pena, Tomas F.

A2 - Cabaleiro, Jose C.

PB - Springer Verlag

ER -

Sodsong W, Mittermayr R, Park Y, Burgstaller B, Blieberger J. Lazy parallel kronecker algebra-operations on heterogeneous multicores. In Rivera FF, Pena TF, Cabaleiro JC, editors, Euro-Par 2017: Parallel Processing - 23rd International Conference on Parallel and Distributed Computing, Proceedings. Springer Verlag. 2017. p. 538-552. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-64203-1_39