Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation

Alan Ayala; Stanimire Tomov; Xi Luo; Hejer Shaeik; Azzam Haidar; George Bosilca; Jack Dongarra

doi:10.1109/ExaMPI49596.2019.00007

Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation

Alan Ayala, Stanimire Tomov, Xi Luo, Hejer Shaeik, Azzam Haidar, George Bosilca, Jack Dongarra

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

21 Scopus citations

Abstract

Most applications targeting exascale, such as those part of the Exascale Computing Project (ECP), are designed for heterogeneous architectures and rely on the Message Passing Interface (MPI) as their underlying parallel programming model. In this paper we analyze the limitations of collective MPI communication for the computation of fast Fourier transforms (FFTs), which are relied on heavily for large-scale particle simulations. We present experiments made at one of the largest heterogeneous platforms, the Summit supercomputer at ORNL. We discuss communication models from state-of-the-art FFT libraries, and propose a new FFT library, named HEFFTE (Highly Efficient FFTs for Exascale), which supports heterogeneous architectures and yields considerable speedups compared with CPU libraries, while maintaining good weak as well as strong scalability.

Original language	English
Title of host publication	Proceedings of ExaMPI 2019
Subtitle of host publication	Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	12-18
Number of pages	7
ISBN (Electronic)	9781728160092
DOIs	https://doi.org/10.1109/ExaMPI49596.2019.00007
State	Published - Nov 2019
Externally published	Yes
Event	2019 IEEE/ACM Workshop on Exascale MPI, ExaMPI 2019 - Denver, United States Duration: Nov 17 2019 → …

Publication series

Name	Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis

Conference

Conference	2019 IEEE/ACM Workshop on Exascale MPI, ExaMPI 2019
Country/Territory	United States
City	Denver
Period	11/17/19 → …

Funding

ACKNOWLEDGMENT This research was supported by the Exascale Computing Project (ECP), Project Number: 17-SC-20-SC, a collaborative effort of two DOE organizations (the Office of Science and the National Nuclear Security Administration) responsible for the planning and preparation of a capable exascale ecosystem.

Access to Document

10.1109/ExaMPI49596.2019.00007

Cite this

Ayala, A., Tomov, S., Luo, X., Shaeik, H., Haidar, A., Bosilca, G., & Dongarra, J. (2019). Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation. In Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis (pp. 12-18). Article 8955469 (Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ExaMPI49596.2019.00007

Ayala, Alan ; Tomov, Stanimire ; Luo, Xi et al. / Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation. Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 12-18 (Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis).

@inproceedings{0c320a76a08440cc82f7f75dd33246f8,

title = "Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation",

abstract = "Most applications targeting exascale, such as those part of the Exascale Computing Project (ECP), are designed for heterogeneous architectures and rely on the Message Passing Interface (MPI) as their underlying parallel programming model. In this paper we analyze the limitations of collective MPI communication for the computation of fast Fourier transforms (FFTs), which are relied on heavily for large-scale particle simulations. We present experiments made at one of the largest heterogeneous platforms, the Summit supercomputer at ORNL. We discuss communication models from state-of-the-art FFT libraries, and propose a new FFT library, named HEFFTE (Highly Efficient FFTs for Exascale), which supports heterogeneous architectures and yields considerable speedups compared with CPU libraries, while maintaining good weak as well as strong scalability.",

author = "Alan Ayala and Stanimire Tomov and Xi Luo and Hejer Shaeik and Azzam Haidar and George Bosilca and Jack Dongarra",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 2019 IEEE/ACM Workshop on Exascale MPI, ExaMPI 2019 ; Conference date: 17-11-2019",

year = "2019",

month = nov,

doi = "10.1109/ExaMPI49596.2019.00007",

language = "English",

series = "Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "12--18",

booktitle = "Proceedings of ExaMPI 2019",

}

Ayala, A, Tomov, S, Luo, X, Shaeik, H, Haidar, A, Bosilca, G & Dongarra, J 2019, Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation. in Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis., 8955469, Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis, Institute of Electrical and Electronics Engineers Inc., pp. 12-18, 2019 IEEE/ACM Workshop on Exascale MPI, ExaMPI 2019, Denver, United States, 11/17/19. https://doi.org/10.1109/ExaMPI49596.2019.00007

Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation. / Ayala, Alan; Tomov, Stanimire; Luo, Xi et al.
Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis. Institute of Electrical and Electronics Engineers Inc., 2019. p. 12-18 8955469 (Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis).

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

TY - GEN

T1 - Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation

AU - Ayala, Alan

AU - Tomov, Stanimire

AU - Luo, Xi

AU - Shaeik, Hejer

AU - Haidar, Azzam

AU - Bosilca, George

AU - Dongarra, Jack

PY - 2019/11

Y1 - 2019/11

N2 - Most applications targeting exascale, such as those part of the Exascale Computing Project (ECP), are designed for heterogeneous architectures and rely on the Message Passing Interface (MPI) as their underlying parallel programming model. In this paper we analyze the limitations of collective MPI communication for the computation of fast Fourier transforms (FFTs), which are relied on heavily for large-scale particle simulations. We present experiments made at one of the largest heterogeneous platforms, the Summit supercomputer at ORNL. We discuss communication models from state-of-the-art FFT libraries, and propose a new FFT library, named HEFFTE (Highly Efficient FFTs for Exascale), which supports heterogeneous architectures and yields considerable speedups compared with CPU libraries, while maintaining good weak as well as strong scalability.

AB - Most applications targeting exascale, such as those part of the Exascale Computing Project (ECP), are designed for heterogeneous architectures and rely on the Message Passing Interface (MPI) as their underlying parallel programming model. In this paper we analyze the limitations of collective MPI communication for the computation of fast Fourier transforms (FFTs), which are relied on heavily for large-scale particle simulations. We present experiments made at one of the largest heterogeneous platforms, the Summit supercomputer at ORNL. We discuss communication models from state-of-the-art FFT libraries, and propose a new FFT library, named HEFFTE (Highly Efficient FFTs for Exascale), which supports heterogeneous architectures and yields considerable speedups compared with CPU libraries, while maintaining good weak as well as strong scalability.

UR - http://www.scopus.com/inward/record.url?scp=85079287951&partnerID=8YFLogxK

U2 - 10.1109/ExaMPI49596.2019.00007

DO - 10.1109/ExaMPI49596.2019.00007

M3 - Conference contribution

AN - SCOPUS:85079287951

T3 - Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis

SP - 12

EP - 18

BT - Proceedings of ExaMPI 2019

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2019 IEEE/ACM Workshop on Exascale MPI, ExaMPI 2019

Y2 - 17 November 2019

ER -

Ayala A, Tomov S, Luo X, Shaeik H, Haidar A, Bosilca G et al. Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation. In Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis. Institute of Electrical and Electronics Engineers Inc. 2019. p. 12-18. 8955469. (Proceedings of ExaMPI 2019: Workshop on Exascale MPI - Held in conjunction with SC 2019: The International Conference for High Performance Computing, Networking, Storage and Analysis). doi: 10.1109/ExaMPI49596.2019.00007

Impacts of Multi-GPU MPI Collective Communications on Large FFT Computation

Abstract

Publication series

Conference

Funding

Access to Document

Other files and links

Fingerprint

Cite this