Topology-oblivious optimization of MPI broadcast algorithms on extreme-scale platforms

Khalid Hasanov; Jean-Noël Quintin; Alexey Lastovetsky

doi:10.1016/j.simpat.2015.03.005

Topology-oblivious optimization of MPI broadcast algorithms on extreme-scale platforms

Khalid Hasanov, Jean-Noël Quintin, Alexey Lastovetsky

Research output: Contribution to journal › Article › peer-review

6 Scopus citations

Abstract

© 2015 Elsevier B.V. All rights reserved. Significant research has been conducted in collective communication operations, in particular in MPI broadcast, on distributed memory platforms. Most of the research efforts aim to optimize the collective operations for particular architectures by taking into account either their topology or platform parameters. In this work we propose a simple but general approach to optimization of the legacy MPI broadcast algorithms, which are widely used in MPICH and Open MPI. The proposed optimization technique is designed to address the challenge of extreme scale of future HPC platforms. It is based on hierarchical transformation of the traditionally flat logical arrangement of communicating processors. Theoretical analysis and experimental results on IBM BlueGene/P and a cluster of the Grid'5000 platform are presented.

Original language	English (US)
Pages (from-to)	30-39
Number of pages	10
Journal	Simulation Modelling Practice and Theory
Volume	58
DOIs	https://doi.org/10.1016/j.simpat.2015.03.005
State	Published - Nov 2015
Externally published	Yes

Bibliographical note

KAUST Repository Item: Exported on 2020-10-01
Acknowledgements: This work has emanated from research conducted with the financial support of IRCSET (Irish Research Council for Science, Engineering and Technology) and IBM, Grant No. EPSPG/2011/188 and Science Foundation Ireland, Grant No. 08/IN.1/I2054.Some of the experiments presented in this publication were carried out using the Grid’5000 experimental testbed, being developed under the INRIA ALADDIN development action with support from CNRS, RENATER and several Universities as well as other funding bodies (see https://www.grid5000.fr).Another part of the experiments were carried out using the resources of the Supercomputing Laboratory at King Abdullah University of Science&Technology (KAUST) in Thuwal, Saudi Arabia.
This publication acknowledges KAUST support, but has no KAUST affiliated authors.

Access to Document

10.1016/j.simpat.2015.03.005

Cite this

@article{8a766778c3a54d5483e90e257706421c,

title = "Topology-oblivious optimization of MPI broadcast algorithms on extreme-scale platforms",

abstract = "{\textcopyright} 2015 Elsevier B.V. All rights reserved. Significant research has been conducted in collective communication operations, in particular in MPI broadcast, on distributed memory platforms. Most of the research efforts aim to optimize the collective operations for particular architectures by taking into account either their topology or platform parameters. In this work we propose a simple but general approach to optimization of the legacy MPI broadcast algorithms, which are widely used in MPICH and Open MPI. The proposed optimization technique is designed to address the challenge of extreme scale of future HPC platforms. It is based on hierarchical transformation of the traditionally flat logical arrangement of communicating processors. Theoretical analysis and experimental results on IBM BlueGene/P and a cluster of the Grid'5000 platform are presented.",

author = "Khalid Hasanov and Jean-No{\"e}l Quintin and Alexey Lastovetsky",

note = "KAUST Repository Item: Exported on 2020-10-01 Acknowledgements: This work has emanated from research conducted with the financial support of IRCSET (Irish Research Council for Science, Engineering and Technology) and IBM, Grant No. EPSPG/2011/188 and Science Foundation Ireland, Grant No. 08/IN.1/I2054.Some of the experiments presented in this publication were carried out using the Grid{\textquoteright}5000 experimental testbed, being developed under the INRIA ALADDIN development action with support from CNRS, RENATER and several Universities as well as other funding bodies (see https://www.grid5000.fr).Another part of the experiments were carried out using the resources of the Supercomputing Laboratory at King Abdullah University of Science&Technology (KAUST) in Thuwal, Saudi Arabia. This publication acknowledges KAUST support, but has no KAUST affiliated authors.",

year = "2015",

month = nov,

doi = "10.1016/j.simpat.2015.03.005",

language = "English (US)",

volume = "58",

pages = "30--39",

journal = "Simulation Modelling Practice and Theory",

issn = "1569-190X",

publisher = "Elsevier B.V.",

}

TY - JOUR

T1 - Topology-oblivious optimization of MPI broadcast algorithms on extreme-scale platforms

AU - Hasanov, Khalid

AU - Quintin, Jean-Noël

AU - Lastovetsky, Alexey

N1 - KAUST Repository Item: Exported on 2020-10-01 Acknowledgements: This work has emanated from research conducted with the financial support of IRCSET (Irish Research Council for Science, Engineering and Technology) and IBM, Grant No. EPSPG/2011/188 and Science Foundation Ireland, Grant No. 08/IN.1/I2054.Some of the experiments presented in this publication were carried out using the Grid’5000 experimental testbed, being developed under the INRIA ALADDIN development action with support from CNRS, RENATER and several Universities as well as other funding bodies (see https://www.grid5000.fr).Another part of the experiments were carried out using the resources of the Supercomputing Laboratory at King Abdullah University of Science&Technology (KAUST) in Thuwal, Saudi Arabia. This publication acknowledges KAUST support, but has no KAUST affiliated authors.

PY - 2015/11

Y1 - 2015/11

N2 - © 2015 Elsevier B.V. All rights reserved. Significant research has been conducted in collective communication operations, in particular in MPI broadcast, on distributed memory platforms. Most of the research efforts aim to optimize the collective operations for particular architectures by taking into account either their topology or platform parameters. In this work we propose a simple but general approach to optimization of the legacy MPI broadcast algorithms, which are widely used in MPICH and Open MPI. The proposed optimization technique is designed to address the challenge of extreme scale of future HPC platforms. It is based on hierarchical transformation of the traditionally flat logical arrangement of communicating processors. Theoretical analysis and experimental results on IBM BlueGene/P and a cluster of the Grid'5000 platform are presented.

AB - © 2015 Elsevier B.V. All rights reserved. Significant research has been conducted in collective communication operations, in particular in MPI broadcast, on distributed memory platforms. Most of the research efforts aim to optimize the collective operations for particular architectures by taking into account either their topology or platform parameters. In this work we propose a simple but general approach to optimization of the legacy MPI broadcast algorithms, which are widely used in MPICH and Open MPI. The proposed optimization technique is designed to address the challenge of extreme scale of future HPC platforms. It is based on hierarchical transformation of the traditionally flat logical arrangement of communicating processors. Theoretical analysis and experimental results on IBM BlueGene/P and a cluster of the Grid'5000 platform are presented.

UR - http://hdl.handle.net/10754/600038

UR - https://linkinghub.elsevier.com/retrieve/pii/S1569190X15000465

UR - http://www.scopus.com/inward/record.url?scp=84944275700&partnerID=8YFLogxK

U2 - 10.1016/j.simpat.2015.03.005

DO - 10.1016/j.simpat.2015.03.005

M3 - Article

SN - 1569-190X

VL - 58

SP - 30

EP - 39

JO - Simulation Modelling Practice and Theory

JF - Simulation Modelling Practice and Theory

ER -

Topology-oblivious optimization of MPI broadcast algorithms on extreme-scale platforms

Abstract

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this