Evaluation of global synchronization for iterative algebra algorithms on many-core

Ayaz ul Hasan Khan; Mayez Al-Mouhamed; Lutfi A. Firdaus

doi:10.1109/SNPD.2015.7176173

Evaluation of global synchronization for iterative algebra algorithms on many-core

Ayaz ul Hasan Khan, Mayez Al-Mouhamed, Lutfi A. Firdaus

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

3 Scopus citations

Abstract

© 2015 IEEE. Massively parallel computing is applied extensively in various scientific and engineering domains. With the growing interest in many-core architectures and due to the lack of explicit support for inter-block synchronization specifically in GPUs, synchronization becomes necessary to minimize inter-block communication time. In this paper, we have proposed two new inter-block synchronization techniques: 1) Relaxed Synchronization, and 2) Block-Query Synchronization. These schemes are used in implementing numerical iterative solvers where computation/communication overlapping is one used optimization to enhance application performance. We have evaluated and analyzed the performance of the proposed synchronization techniques using Jacobi Iterative Solver in comparison to the state of the art inter-block lock-free synchronization techniques. We have achieved about 1-8% performance improvement in terms of execution time over lock-free synchronization depending on the problem size and the number of thread blocks. We have also evaluated the proposed algorithm on GPU and MIC architectures and obtained about 8-26% performance improvement over the barrier synchronization available in OpenMP programming environment depending on the problem size and number of cores used.

Original language	English (US)
Title of host publication	2015 IEEE/ACIS 16th International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD)
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
ISBN (Print)	9781479986767
DOIs	https://doi.org/10.1109/SNPD.2015.7176173
State	Published - Jun 2015
Externally published	Yes

Bibliographical note

KAUST Repository Item: Exported on 2020-10-01
Acknowledgements: This project was funded by the National Plan for Science, Technology and Innovation (MAARIFAH) – King Abdulaziz City for Science and Technology – through the Science and Technology Unit at King Fahd University of Petroleum & Minerals (KFUPM) – the Kingdom of Saudi Arabia, award number (12-INF3008-04). Thanks to King Abdullah University of Science and Technology (KAUST) for providing access to its K20X GPU cluster.
This publication acknowledges KAUST support, but has no KAUST affiliated authors.

Access to Document

10.1109/SNPD.2015.7176173

Cite this

ul Hasan Khan, A., Al-Mouhamed, M., & Firdaus, L. A. (2015). Evaluation of global synchronization for iterative algebra algorithms on many-core. In 2015 IEEE/ACIS 16th International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD) Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/SNPD.2015.7176173

@inproceedings{b067a79bd3eb4a2ca5bdc39a02e18db2,

title = "Evaluation of global synchronization for iterative algebra algorithms on many-core",

abstract = "{\textcopyright} 2015 IEEE. Massively parallel computing is applied extensively in various scientific and engineering domains. With the growing interest in many-core architectures and due to the lack of explicit support for inter-block synchronization specifically in GPUs, synchronization becomes necessary to minimize inter-block communication time. In this paper, we have proposed two new inter-block synchronization techniques: 1) Relaxed Synchronization, and 2) Block-Query Synchronization. These schemes are used in implementing numerical iterative solvers where computation/communication overlapping is one used optimization to enhance application performance. We have evaluated and analyzed the performance of the proposed synchronization techniques using Jacobi Iterative Solver in comparison to the state of the art inter-block lock-free synchronization techniques. We have achieved about 1-8% performance improvement in terms of execution time over lock-free synchronization depending on the problem size and the number of thread blocks. We have also evaluated the proposed algorithm on GPU and MIC architectures and obtained about 8-26% performance improvement over the barrier synchronization available in OpenMP programming environment depending on the problem size and number of cores used.",

author = "{ul Hasan Khan}, Ayaz and Mayez Al-Mouhamed and Firdaus, {Lutfi A.}",

note = "KAUST Repository Item: Exported on 2020-10-01 Acknowledgements: This project was funded by the National Plan for Science, Technology and Innovation (MAARIFAH) – King Abdulaziz City for Science and Technology – through the Science and Technology Unit at King Fahd University of Petroleum & Minerals (KFUPM) – the Kingdom of Saudi Arabia, award number (12-INF3008-04). Thanks to King Abdullah University of Science and Technology (KAUST) for providing access to its K20X GPU cluster. This publication acknowledges KAUST support, but has no KAUST affiliated authors.",

year = "2015",

month = jun,

doi = "10.1109/SNPD.2015.7176173",

language = "English (US)",

isbn = "9781479986767",

booktitle = "2015 IEEE/ACIS 16th International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD)",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

}

ul Hasan Khan, A, Al-Mouhamed, M & Firdaus, LA 2015, Evaluation of global synchronization for iterative algebra algorithms on many-core. in 2015 IEEE/ACIS 16th International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/SNPD.2015.7176173

Evaluation of global synchronization for iterative algebra algorithms on many-core. / ul Hasan Khan, Ayaz; Al-Mouhamed, Mayez; Firdaus, Lutfi A.
2015 IEEE/ACIS 16th International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD). Institute of Electrical and Electronics Engineers (IEEE), 2015.

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

TY - GEN

T1 - Evaluation of global synchronization for iterative algebra algorithms on many-core

AU - ul Hasan Khan, Ayaz

AU - Al-Mouhamed, Mayez

AU - Firdaus, Lutfi A.

N1 - KAUST Repository Item: Exported on 2020-10-01 Acknowledgements: This project was funded by the National Plan for Science, Technology and Innovation (MAARIFAH) – King Abdulaziz City for Science and Technology – through the Science and Technology Unit at King Fahd University of Petroleum & Minerals (KFUPM) – the Kingdom of Saudi Arabia, award number (12-INF3008-04). Thanks to King Abdullah University of Science and Technology (KAUST) for providing access to its K20X GPU cluster. This publication acknowledges KAUST support, but has no KAUST affiliated authors.

PY - 2015/6

Y1 - 2015/6

N2 - © 2015 IEEE. Massively parallel computing is applied extensively in various scientific and engineering domains. With the growing interest in many-core architectures and due to the lack of explicit support for inter-block synchronization specifically in GPUs, synchronization becomes necessary to minimize inter-block communication time. In this paper, we have proposed two new inter-block synchronization techniques: 1) Relaxed Synchronization, and 2) Block-Query Synchronization. These schemes are used in implementing numerical iterative solvers where computation/communication overlapping is one used optimization to enhance application performance. We have evaluated and analyzed the performance of the proposed synchronization techniques using Jacobi Iterative Solver in comparison to the state of the art inter-block lock-free synchronization techniques. We have achieved about 1-8% performance improvement in terms of execution time over lock-free synchronization depending on the problem size and the number of thread blocks. We have also evaluated the proposed algorithm on GPU and MIC architectures and obtained about 8-26% performance improvement over the barrier synchronization available in OpenMP programming environment depending on the problem size and number of cores used.

AB - © 2015 IEEE. Massively parallel computing is applied extensively in various scientific and engineering domains. With the growing interest in many-core architectures and due to the lack of explicit support for inter-block synchronization specifically in GPUs, synchronization becomes necessary to minimize inter-block communication time. In this paper, we have proposed two new inter-block synchronization techniques: 1) Relaxed Synchronization, and 2) Block-Query Synchronization. These schemes are used in implementing numerical iterative solvers where computation/communication overlapping is one used optimization to enhance application performance. We have evaluated and analyzed the performance of the proposed synchronization techniques using Jacobi Iterative Solver in comparison to the state of the art inter-block lock-free synchronization techniques. We have achieved about 1-8% performance improvement in terms of execution time over lock-free synchronization depending on the problem size and the number of thread blocks. We have also evaluated the proposed algorithm on GPU and MIC architectures and obtained about 8-26% performance improvement over the barrier synchronization available in OpenMP programming environment depending on the problem size and number of cores used.

UR - http://hdl.handle.net/10754/598248

UR - http://ieeexplore.ieee.org/document/7176173/

UR - http://www.scopus.com/inward/record.url?scp=84947104373&partnerID=8YFLogxK

U2 - 10.1109/SNPD.2015.7176173

DO - 10.1109/SNPD.2015.7176173

M3 - Conference contribution

SN - 9781479986767

BT - 2015 IEEE/ACIS 16th International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD)

PB - Institute of Electrical and Electronics Engineers (IEEE)

ER -

ul Hasan Khan A, Al-Mouhamed M, Firdaus LA. Evaluation of global synchronization for iterative algebra algorithms on many-core. In 2015 IEEE/ACIS 16th International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing (SNPD). Institute of Electrical and Electronics Engineers (IEEE). 2015 doi: 10.1109/SNPD.2015.7176173

Evaluation of global synchronization for iterative algebra algorithms on many-core

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this