Multilevel Parallelization: Grid Methods for Solving Direct and Inverse Problems

Sofya Titarenko; Igor Kulikov; Igor Chernykh; Maxim A. Shishlenin; Olga Krivorot'ko; Dmitry A. Voronov; Mark Hildyard

doi:10.1007/978-3-319-55669-7_10

Multilevel Parallelization: Grid Methods for Solving Direct and Inverse Problems

Sofya Titarenko, Igor Kulikov, Igor Chernykh, Maxim A. Shishlenin, Olga Krivorot'ko, Dmitry A. Voronov, Mark Hildyard

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

5 Citations (Scopus)

Abstract

In this paper we present grid methods which we have developed for solving direct and inverse problems, and their realization with different levels of optimization. We have focused on solving systems of hyperbolic equations using finite difference and finite volume numerical methods on multicore architectures. Several levels of parallelism have been applied: geometric decomposition of the calculative domain, workload distribution over threads within OpenMP directives, and vectorization. The run-time efficiency of these methods has been investigated. These developments have been tested using the astrophysics code AstroPhi on a hybrid cluster Polytechnic RSC PetaStream (consisting of Intel Xeon Phi accelerators) and a geophysics (seismic wave) code on an Intel Core i7-3930K multicore processor. We present the results of the calculations and study MPI run-time energy efficiency.

Original language	English
Title of host publication	Supercomputing
Subtitle of host publication	Second Russian Supercomputing Days, RuSCDays 2016, Moscow, Russia, September 26–27, 2016, Revised Selected Papers
Editors	Vladimir Voevodin, Sergey Sobolev
Place of Publication	Cham
Publisher	Springer International Publishing AG
Pages	118-131
Number of pages	14
Edition	1st
ISBN (Electronic)	9783319556697
ISBN (Print)	9783319556680
DOIs	https://doi.org/10.1007/978-3-319-55669-7_10
Publication status	Published - 12 Mar 2017
Externally published	Yes
Event	2nd Russian Supercomputing Days 2016 - Moscow, Russian Federation Duration: 26 Sep 2016 → 27 Sep 2016 Conference number: 2

Publication series

Name	Communications in Computer and Information Science
Publisher	Springer International Publishing AG
ISSN (Print)	1865-0929

Conference

Conference	2nd Russian Supercomputing Days 2016
Abbreviated title	RuSCDays 2016
Country/Territory	Russian Federation
City	Moscow
Period	26/09/16 → 27/09/16

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1007/978-3-319-55669-7_10

Cite this

Titarenko, S., Kulikov, I., Chernykh, I., Shishlenin, M. A., Krivorot'ko, O., Voronov, D. A., & Hildyard, M. (2017). Multilevel Parallelization: Grid Methods for Solving Direct and Inverse Problems. In V. Voevodin, & S. Sobolev (Eds.), Supercomputing: Second Russian Supercomputing Days, RuSCDays 2016, Moscow, Russia, September 26–27, 2016, Revised Selected Papers (1st ed., pp. 118-131). (Communications in Computer and Information Science). Springer International Publishing AG. https://doi.org/10.1007/978-3-319-55669-7_10

Titarenko, Sofya ; Kulikov, Igor ; Chernykh, Igor et al. / Multilevel Parallelization : Grid Methods for Solving Direct and Inverse Problems. Supercomputing: Second Russian Supercomputing Days, RuSCDays 2016, Moscow, Russia, September 26–27, 2016, Revised Selected Papers . editor / Vladimir Voevodin ; Sergey Sobolev. 1st. ed. Cham : Springer International Publishing AG, 2017. pp. 118-131 (Communications in Computer and Information Science).

@inproceedings{25322bac48754234930bb5dedaac6916,

title = "Multilevel Parallelization: Grid Methods for Solving Direct and Inverse Problems",

abstract = "In this paper we present grid methods which we have developed for solving direct and inverse problems, and their realization with different levels of optimization. We have focused on solving systems of hyperbolic equations using finite difference and finite volume numerical methods on multicore architectures. Several levels of parallelism have been applied: geometric decomposition of the calculative domain, workload distribution over threads within OpenMP directives, and vectorization. The run-time efficiency of these methods has been investigated. These developments have been tested using the astrophysics code AstroPhi on a hybrid cluster Polytechnic RSC PetaStream (consisting of Intel Xeon Phi accelerators) and a geophysics (seismic wave) code on an Intel Core i7-3930K multicore processor. We present the results of the calculations and study MPI run-time energy efficiency.",

keywords = "High performance computing, Intel Xeon Phi accelerators, Grid-based numerical methods",

author = "Sofya Titarenko and Igor Kulikov and Igor Chernykh and Shishlenin, {Maxim A.} and Olga Krivorot'ko and Voronov, {Dmitry A.} and Mark Hildyard",

year = "2017",

month = mar,

day = "12",

doi = "10.1007/978-3-319-55669-7_10",

language = "English",

isbn = "9783319556680",

series = "Communications in Computer and Information Science",

publisher = "Springer International Publishing AG",

pages = "118--131",

editor = "Vladimir Voevodin and Sergey Sobolev",

booktitle = "Supercomputing",

address = "Switzerland",

edition = "1st",

note = "2nd Russian Supercomputing Days 2016, RuSCDays 2016 ; Conference date: 26-09-2016 Through 27-09-2016",

}

Titarenko, S, Kulikov, I, Chernykh, I, Shishlenin, MA, Krivorot'ko, O, Voronov, DA & Hildyard, M 2017, Multilevel Parallelization: Grid Methods for Solving Direct and Inverse Problems. in V Voevodin & S Sobolev (eds), Supercomputing: Second Russian Supercomputing Days, RuSCDays 2016, Moscow, Russia, September 26–27, 2016, Revised Selected Papers . 1st edn, Communications in Computer and Information Science, Springer International Publishing AG, Cham, pp. 118-131, 2nd Russian Supercomputing Days 2016, Moscow, Russian Federation, 26/09/16. https://doi.org/10.1007/978-3-319-55669-7_10

Multilevel Parallelization : Grid Methods for Solving Direct and Inverse Problems. / Titarenko, Sofya; Kulikov, Igor; Chernykh, Igor et al.

Supercomputing: Second Russian Supercomputing Days, RuSCDays 2016, Moscow, Russia, September 26–27, 2016, Revised Selected Papers . ed. / Vladimir Voevodin; Sergey Sobolev. 1st. ed. Cham : Springer International Publishing AG, 2017. p. 118-131 (Communications in Computer and Information Science).

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

TY - GEN

T1 - Multilevel Parallelization

T2 - 2nd Russian Supercomputing Days 2016

AU - Titarenko, Sofya

AU - Kulikov, Igor

AU - Chernykh, Igor

AU - Shishlenin, Maxim A.

AU - Krivorot'ko, Olga

AU - Voronov, Dmitry A.

AU - Hildyard, Mark

N1 - Conference code: 2

PY - 2017/3/12

Y1 - 2017/3/12

N2 - In this paper we present grid methods which we have developed for solving direct and inverse problems, and their realization with different levels of optimization. We have focused on solving systems of hyperbolic equations using finite difference and finite volume numerical methods on multicore architectures. Several levels of parallelism have been applied: geometric decomposition of the calculative domain, workload distribution over threads within OpenMP directives, and vectorization. The run-time efficiency of these methods has been investigated. These developments have been tested using the astrophysics code AstroPhi on a hybrid cluster Polytechnic RSC PetaStream (consisting of Intel Xeon Phi accelerators) and a geophysics (seismic wave) code on an Intel Core i7-3930K multicore processor. We present the results of the calculations and study MPI run-time energy efficiency.

AB - In this paper we present grid methods which we have developed for solving direct and inverse problems, and their realization with different levels of optimization. We have focused on solving systems of hyperbolic equations using finite difference and finite volume numerical methods on multicore architectures. Several levels of parallelism have been applied: geometric decomposition of the calculative domain, workload distribution over threads within OpenMP directives, and vectorization. The run-time efficiency of these methods has been investigated. These developments have been tested using the astrophysics code AstroPhi on a hybrid cluster Polytechnic RSC PetaStream (consisting of Intel Xeon Phi accelerators) and a geophysics (seismic wave) code on an Intel Core i7-3930K multicore processor. We present the results of the calculations and study MPI run-time energy efficiency.

KW - High performance computing

KW - Intel Xeon Phi accelerators

KW - Grid-based numerical methods

UR - https://www.scopus.com/record/display.uri?eid=2-s2.0-85015824462&origin=resultslist&sort=plf-f&src=s&st1=Multilevel+parallelization%3a+grid+methods+for+solving+direct+and+inverse+problems&st2=&sid=9207d5478cf0e554b2b0213537909c42&sot=b&sdt=b&sl=95&s=TITLE-ABS-KEY%28Multilevel+parallelization%3a+grid+methods+for+solving+direct+and+inverse+problems%29&relpos=0&citeCnt=0&searchTerm=

U2 - 10.1007/978-3-319-55669-7_10

DO - 10.1007/978-3-319-55669-7_10

M3 - Conference contribution

SN - 9783319556680

T3 - Communications in Computer and Information Science

SP - 118

EP - 131

BT - Supercomputing

A2 - Voevodin, Vladimir

A2 - Sobolev, Sergey

PB - Springer International Publishing AG

CY - Cham

Y2 - 26 September 2016 through 27 September 2016

ER -

Titarenko S, Kulikov I, Chernykh I, Shishlenin MA, Krivorot'ko O, Voronov DA et al. Multilevel Parallelization: Grid Methods for Solving Direct and Inverse Problems. In Voevodin V, Sobolev S, editors, Supercomputing: Second Russian Supercomputing Days, RuSCDays 2016, Moscow, Russia, September 26–27, 2016, Revised Selected Papers . 1st ed. Cham: Springer International Publishing AG. 2017. p. 118-131. (Communications in Computer and Information Science). doi: 10.1007/978-3-319-55669-7_10

Multilevel Parallelization: Grid Methods for Solving Direct and Inverse Problems

Abstract

Publication series

Conference

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this