3D multi-source CSEM simulations: Feasibility and comparison of parallel direct solvers

Vladimir Puzyrev, Seid Koric

Research output: Contribution to journalConference article

Abstract

Modern numerical algorithms for computational electromagnetics lead to many large sparse systems of linear equations. Their solution takes up to 90% of the total computational time in the geophysical inversion process. This paper provides evaluation and comparison of several state-of-the-art direct solvers in a massively parallel environment. We determine the largest complex systems that can be solved today with these methods and evaluate their performance and scalability on one of the world's most powerful supercomputers. Small sensitivity of direct methods to the number of sources, modeling frequency and conductivity distribution in the subsurface is confirmed. The results show the potentials and limitations of different parallel implementations on a petascale high-performance computing system.

Original languageEnglish (US)
Pages (from-to)833-838
Number of pages6
JournalSEG Technical Program Expanded Abstracts
Volume34
DOIs
StatePublished - Jan 1 2015
EventSEG New Orleans Annual Meeting, SEG 2015 - New Orleans, United States
Duration: Oct 18 2011Oct 23 2011

Fingerprint

Computational electromagnetics
Supercomputers
Linear equations
Scalability
Large scale systems
computational electromagnetics
supercomputers
frequency distribution
linear equations
complex systems
simulation
conductivity
inversions
evaluation
sensitivity
modeling
comparison
method
distribution
inversion

ASJC Scopus subject areas

  • Geotechnical Engineering and Engineering Geology
  • Geophysics

Cite this

3D multi-source CSEM simulations : Feasibility and comparison of parallel direct solvers. / Puzyrev, Vladimir; Koric, Seid.

In: SEG Technical Program Expanded Abstracts, Vol. 34, 01.01.2015, p. 833-838.

Research output: Contribution to journalConference article

@article{39fc223488dd47ae8b98575fc74024a5,
title = "3D multi-source CSEM simulations: Feasibility and comparison of parallel direct solvers",
abstract = "Modern numerical algorithms for computational electromagnetics lead to many large sparse systems of linear equations. Their solution takes up to 90{\%} of the total computational time in the geophysical inversion process. This paper provides evaluation and comparison of several state-of-the-art direct solvers in a massively parallel environment. We determine the largest complex systems that can be solved today with these methods and evaluate their performance and scalability on one of the world's most powerful supercomputers. Small sensitivity of direct methods to the number of sources, modeling frequency and conductivity distribution in the subsurface is confirmed. The results show the potentials and limitations of different parallel implementations on a petascale high-performance computing system.",
author = "Vladimir Puzyrev and Seid Koric",
year = "2015",
month = "1",
day = "1",
doi = "10.1190/segam2015-5864005.1",
language = "English (US)",
volume = "34",
pages = "833--838",
journal = "SEG Technical Program Expanded Abstracts",
issn = "1052-3812",
publisher = "Society of Exploration Geophysicists",

}

TY - JOUR

T1 - 3D multi-source CSEM simulations

T2 - Feasibility and comparison of parallel direct solvers

AU - Puzyrev, Vladimir

AU - Koric, Seid

PY - 2015/1/1

Y1 - 2015/1/1

N2 - Modern numerical algorithms for computational electromagnetics lead to many large sparse systems of linear equations. Their solution takes up to 90% of the total computational time in the geophysical inversion process. This paper provides evaluation and comparison of several state-of-the-art direct solvers in a massively parallel environment. We determine the largest complex systems that can be solved today with these methods and evaluate their performance and scalability on one of the world's most powerful supercomputers. Small sensitivity of direct methods to the number of sources, modeling frequency and conductivity distribution in the subsurface is confirmed. The results show the potentials and limitations of different parallel implementations on a petascale high-performance computing system.

AB - Modern numerical algorithms for computational electromagnetics lead to many large sparse systems of linear equations. Their solution takes up to 90% of the total computational time in the geophysical inversion process. This paper provides evaluation and comparison of several state-of-the-art direct solvers in a massively parallel environment. We determine the largest complex systems that can be solved today with these methods and evaluate their performance and scalability on one of the world's most powerful supercomputers. Small sensitivity of direct methods to the number of sources, modeling frequency and conductivity distribution in the subsurface is confirmed. The results show the potentials and limitations of different parallel implementations on a petascale high-performance computing system.

UR - http://www.scopus.com/inward/record.url?scp=84997488000&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84997488000&partnerID=8YFLogxK

U2 - 10.1190/segam2015-5864005.1

DO - 10.1190/segam2015-5864005.1

M3 - Conference article

AN - SCOPUS:84997488000

VL - 34

SP - 833

EP - 838

JO - SEG Technical Program Expanded Abstracts

JF - SEG Technical Program Expanded Abstracts

SN - 1052-3812

ER -