Cpp-Taskflow: Fast Task-Based Parallel Programming Using Modern C++

Tsung-Wei Huang; Chun-Xun Lin; Guannan Guo; Martin Wong

doi:10.1109/IPDPS.2019.00105

Cpp-Taskflow: Fast Task-Based Parallel Programming Using Modern C++

Tsung-Wei Huang, Chun-Xun Lin, Guannan Guo, Martin Wong

Office of the Provost

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

85 Citations (Scopus)

Abstract

In this paper we introduce Cpp-Taskflow, a new C++ tasking library to help developers quickly write parallel programs using task dependency graphs. Cpp-Taskflow leverages the power of modern C++ and task-based approaches to enable efficient implementations of parallel decomposition strategies. Our programming model can quickly handle not only traditional loop-level parallelism, but also irregular patterns such as graph algorithms, incremental flows, and dynamic data structures. Compared with existing libraries, Cpp-Taskflow is more cost efficient in performance scaling and software integration. We have evaluated Cpp-Taskflow on both micro-benchmarks and real-world applications with million-scale tasking. In a machine learning example, Cpp-Taskflow achieved 1.5-2.7× less coding complexity and 14-38% speed-up over two industrial-strength libraries OpenMP Tasking and Intel Threading Building Blocks (TBB).

Original language	English
Title of host publication	2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS)
Publisher	IEEE
Pages	974-983
Number of pages	10
ISBN (Electronic)	9781728112466
ISBN (Print)	9781728112473 (Print on Demand)
DOIs	https://doi.org/10.1109/IPDPS.2019.00105
Publication status	Published - May 2019
Event	2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS) - Rio de Janeiro, Brazil Duration: 20 May 2019 → 24 May 2019 https://ieeexplore.ieee.org/xpl/conhome/8804711/proceeding (Link to conference proceedings)

Publication series

Name	Proceedings of IEEE International Parallel and Distributed Processing Symposium (IPDPS)

Symposium

Symposium	2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS)
Country/Territory	Brazil
City	Rio de Janeiro
Period	20/05/19 → 24/05/19
Internet address	https://ieeexplore.ieee.org/xpl/conhome/8804711/proceeding (Link to conference proceedings)

User-Defined Keywords

Task analysis
parallel programming library and model
multithreading

Access to Document

10.1109/IPDPS.2019.00105

Cite this

Huang, T.-W., Lin, C.-X., Guo, G., & Wong, M. (2019). Cpp-Taskflow: Fast Task-Based Parallel Programming Using Modern C++. In 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS) (pp. 974-983). Article 8821011 (Proceedings of IEEE International Parallel and Distributed Processing Symposium (IPDPS)). IEEE. https://doi.org/10.1109/IPDPS.2019.00105

@inproceedings{bf82314c6b044a64970af1f8605e088e,

title = "Cpp-Taskflow: Fast Task-Based Parallel Programming Using Modern C++",

abstract = "In this paper we introduce Cpp-Taskflow, a new C++ tasking library to help developers quickly write parallel programs using task dependency graphs. Cpp-Taskflow leverages the power of modern C++ and task-based approaches to enable efficient implementations of parallel decomposition strategies. Our programming model can quickly handle not only traditional loop-level parallelism, but also irregular patterns such as graph algorithms, incremental flows, and dynamic data structures. Compared with existing libraries, Cpp-Taskflow is more cost efficient in performance scaling and software integration. We have evaluated Cpp-Taskflow on both micro-benchmarks and real-world applications with million-scale tasking. In a machine learning example, Cpp-Taskflow achieved 1.5-2.7× less coding complexity and 14-38% speed-up over two industrial-strength libraries OpenMP Tasking and Intel Threading Building Blocks (TBB).",

keywords = "Task analysis, parallel programming library and model, multithreading",

author = "Tsung-Wei Huang and Chun-Xun Lin and Guannan Guo and Martin Wong",

note = "Funding information: This work is supported by NSF Grant CCF-1718883 and DARPA Grant FA-8650-18-2-7843. ; 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS) ; Conference date: 20-05-2019 Through 24-05-2019",

year = "2019",

month = may,

doi = "10.1109/IPDPS.2019.00105",

language = "English",

isbn = "9781728112473 (Print on Demand)",

series = "Proceedings of IEEE International Parallel and Distributed Processing Symposium (IPDPS)",

publisher = "IEEE",

pages = "974--983",

booktitle = "2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS)",

address = "United States",

url = "https://ieeexplore.ieee.org/xpl/conhome/8804711/proceeding",

}

Huang, T-W, Lin, C-X, Guo, G & Wong, M 2019, Cpp-Taskflow: Fast Task-Based Parallel Programming Using Modern C++. in 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS)., 8821011, Proceedings of IEEE International Parallel and Distributed Processing Symposium (IPDPS), IEEE, pp. 974-983, 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS), Rio de Janeiro, Brazil, 20/05/19. https://doi.org/10.1109/IPDPS.2019.00105

Cpp-Taskflow: Fast Task-Based Parallel Programming Using Modern C++. / Huang, Tsung-Wei; Lin, Chun-Xun; Guo, Guannan et al.
2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS). IEEE, 2019. p. 974-983 8821011 (Proceedings of IEEE International Parallel and Distributed Processing Symposium (IPDPS)).

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

TY - GEN

T1 - Cpp-Taskflow: Fast Task-Based Parallel Programming Using Modern C++

AU - Huang, Tsung-Wei

AU - Lin, Chun-Xun

AU - Guo, Guannan

AU - Wong, Martin

N1 - Funding information: This work is supported by NSF Grant CCF-1718883 and DARPA Grant FA-8650-18-2-7843.

PY - 2019/5

Y1 - 2019/5

N2 - In this paper we introduce Cpp-Taskflow, a new C++ tasking library to help developers quickly write parallel programs using task dependency graphs. Cpp-Taskflow leverages the power of modern C++ and task-based approaches to enable efficient implementations of parallel decomposition strategies. Our programming model can quickly handle not only traditional loop-level parallelism, but also irregular patterns such as graph algorithms, incremental flows, and dynamic data structures. Compared with existing libraries, Cpp-Taskflow is more cost efficient in performance scaling and software integration. We have evaluated Cpp-Taskflow on both micro-benchmarks and real-world applications with million-scale tasking. In a machine learning example, Cpp-Taskflow achieved 1.5-2.7× less coding complexity and 14-38% speed-up over two industrial-strength libraries OpenMP Tasking and Intel Threading Building Blocks (TBB).

AB - In this paper we introduce Cpp-Taskflow, a new C++ tasking library to help developers quickly write parallel programs using task dependency graphs. Cpp-Taskflow leverages the power of modern C++ and task-based approaches to enable efficient implementations of parallel decomposition strategies. Our programming model can quickly handle not only traditional loop-level parallelism, but also irregular patterns such as graph algorithms, incremental flows, and dynamic data structures. Compared with existing libraries, Cpp-Taskflow is more cost efficient in performance scaling and software integration. We have evaluated Cpp-Taskflow on both micro-benchmarks and real-world applications with million-scale tasking. In a machine learning example, Cpp-Taskflow achieved 1.5-2.7× less coding complexity and 14-38% speed-up over two industrial-strength libraries OpenMP Tasking and Intel Threading Building Blocks (TBB).

KW - Task analysis

KW - parallel programming library and model

KW - multithreading

UR - https://www.scopus.com/record/display.uri?eid=2-s2.0-85067794112&origin=inward

U2 - 10.1109/IPDPS.2019.00105

DO - 10.1109/IPDPS.2019.00105

M3 - Conference proceeding

SN - 9781728112473 (Print on Demand)

T3 - Proceedings of IEEE International Parallel and Distributed Processing Symposium (IPDPS)

SP - 974

EP - 983

BT - 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS)

PB - IEEE

T2 - 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS)

Y2 - 20 May 2019 through 24 May 2019

ER -

Cpp-Taskflow: Fast Task-Based Parallel Programming Using Modern C++

Abstract

Publication series

Symposium

User-Defined Keywords

Access to Document

Other files and links

Fingerprint

Cite this