Pasar al contenido principal

ALGORITMOS PARALELOS Y SOFTWARE PARA METODOS ALGEBRAICOS EN ANALISIS DE DATOS

PID2019-107379RB-I00

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal de I+D+i Orientado a los Retos de la Sociedad
Convocatoria Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Año convocatoria 2019
Plan Estatal 2017-2020 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSITAT POLITÈCNICA DE VALÈNCIA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 14
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Physically agnostic quasi normal mode expansion in time dispersive structures: From mechanical vibrations to nanophotonic resonances

  • Nicolet, André
  • Demesy, Guillaume
  • Zolla, Frédéric
  • Campos, Carmen
  • Geuzaine, Christophe
  • Jose E. Roman|||0000-0003-1144-6772
[EN] Resonances, also known as quasi normal modes (QNM) in the non-Hermitian case, play an ubiquitous role in all domains of physics ruled by wave phenomena, notably in continuum mechanics, acoustics, electrodynamics, and quantum theory. In this paper, we present a QNM expansion for dispersive systems, recently applied to photonics but based on sixty year old techniques in mechanics. The resulting numerical algorithm appears to be physically agnostic, that is independent of the considered physical problem and can therefore be implemented as a mere toolbox in a nonlinear eigenvalue computation library., Jose E. Roman was partially supported by the Spanish Agencia Estatal de Investigacion under grant PID2019-107379RB-I00/AEI/10.13039/501100011033.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

NEP: A Module for the Parallel Solution of Nonlinear Eigenvalue Problems in SLEPc

  • Campos, Carmen
  • Jose E. Roman|||0000-0003-1144-6772
[EN] SLEPc is a parallel library for the solution of various types of large-scale eigenvalue problems. Over the past few years, we have been developing a module within SLEPc, called NEP, that is intended for solving nonlinear eigenvalue problems. These problems can be defined by means of a matrix-valued function that depends nonlinearly on a single scalar parameter. We do not consider the particular case of polynomial eigenvalue problems (which are implemented in a different module in SLEPc) and focus here on rational eigenvalue problems and other general nonlinear eigenproblems involving square roots or any other nonlinear function. The article discusses how the NEP module has been designed to fit the needs of applications and provides a description of the available solvers, including some implementation details such as parallelization. Several test problems coming from real applications are used to evaluate the performance and reliability of the solvers., This work was partially funded by the Spanish Agencia Estatal de Investigacion AEI http://ciencia.gob.es under grants TIN2016-75985-P AEI and PID2019-107379RB-I00 AEI (including European Commission FEDER funds).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A Parallel Structured Divide-and-Conquer Algorithm for Symmetric Tridiagonal Eigenvalue Problems

  • Liao, Xia
  • Li, Shengguo
  • Lu, Yutong
  • Jose E. Roman|||0000-0003-1144-6772
[EN] In this article, a parallel structured divide-and-conquer (PSDC) eigensolver is proposed for symmetric tridiagonal matrices based on ScaLAPACK and a parallel structured matrix multiplication algorithm, called PSMMA. Computing the eigenvectors via matrix-matrix multiplications is the most computationally expensive part of the divide-and-conquer algorithm, and one of the matrices involved in such multiplications is a rank-structured Cauchy-like matrix. By exploiting this particular property, PSMMA constructs the local matrices by using generators of Cauchy-like matrices without any communication, and further reduces the computation costs by using a structured low-rank approximation algorithm. Thus, both the communication and computation costs are reduced. Experimental results show that both PSMMA and PSDC are highly scalable and scale to 4096 processes at least. PSDC has better scalability than PHDC that was proposed in [16] and only scaled to 300 processes for the same matrices. Comparing with PDSTEDC in ScaLAPACK, PSDC is always faster and achieves 1.4x-1.6x speedup for some matrices with few deflations. PSDC is also comparable with ELPA, with PSDC being faster than ELPA when using few processes and a little slower when using many processes., The authors would like to thank the referees for their valuable comments which greatly improve the presentation of this article. This work was supported by National Natural Science Foundation of China (No. NNW2019ZT6-B20, NNW2019ZT6B21, NNW2019ZT5-A10, U1611261, 61872392, and U1811461), National Key RD Program of China (2018YFB0204303), NSF of Hunan (No. 2019JJ40339), NSF of NUDT (No. ZK18-03-01), Guangdong Natural Science Foundation (2018B030312002), and the Program for Guangdong Introducing Innovative and Entrepreneurial Teams under Grant 2016ZT06D211. The work of Jose E. Roman was supported by the Spanish Agencia Estatal de Investigacion (AEI) under project SLEPc-DA (PID2019-107379RB-I00).
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Improvements to SLEPc in releases 3.14-3.18

  • Jose E. Roman|||0000-0003-1144-6772
  • Alvarruiz Bermejo, Fernando|||0000-0001-5957-9561
  • Lamas Daviña, Alejandro
  • Campos, Carmen
  • Dalcin, Lisandro
  • Jolivet, Pierre
[EN] This short article describes the main newfeatures added to SLEPc, the Scalable Library for Eigenvalue Problem Computations, in the past two and a half years, corresponding to five release versions. The main novelty is the extension of the SVD module with new problem types, such as the generalized SVD or the hyperbolic SVD. Additionally, many improvements have been incorporated in different parts of the library, including contour integral eigensolvers, preconditioning, and GPU support.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Solving differential eigenproblems via the spectral Tau method

  • Vasconcelos, P. B.
  • Matos, J. M. A.
  • Jose E. Roman|||0000-0003-1144-6772
[EN] The spectral Tau method to compute eigenpairs of ordinary differential equations is implemented as part of the Tau Toolbox-a numerical library for the solution of integro-differential problems. This mathematical software enables a symbolic syntax to be applied to objects to manipulate and solve differential problems with ease and accuracy. The library is explained in detail and its application to various problems is illustrated: numerical approximations for linear, quadratic, and nonlinear differential eigenvalue problems., This work was partially supported by the Spanish Agencia Estatal de Investigacion under grant PID2019-107379RB-I00/AEI/10.13039/501100011033, and by CMUP, which is financed by national funds through FCT-Fundacao para a Ciencia e a Tecnologia, I.P., under the project with reference UIDB/00144/2020.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A parallel structured banded DC algorithm for symmetric eigenvalue problems

  • Li, Shengguo
  • Liao, Xia
  • Lu, Yutong
  • Yue, Xiaoqiang
  • Jose E. Roman|||0000-0003-1144-6772
[EN] In this paper, a novel parallel structured divide-and-conquer (DC) algorithm is proposed for symmetric banded eigenvalue problems, denoted by PBSDC, which modifes the classical parallel banded DC (PBDC) algorithm by reducing its computational cost. The main tool that PBSDC uses is a parallel structured matrix multiplication algorithm (PSMMA), which can be much faster than the general dense matrix multiplication ScaLAPACK routine PDGEMM. Numerous experiments have been performed on Tianhe-2 supercomputer to compare PBSDC with PBDC and ELPA. For matrices with few defations, PBSDC can be much faster than PBDC since computations are saved. For matrices with many defations and/or small bandwidths, PBSDC can be faster than the tridiagonalization-based DC implemented in LAPACK and ELPA. However, PBSDC would become slower than ELPA for matrices with relatively large bandwidths., The authors would like to thank the referees for their valuable comments. This work is supported in part by NSFC (No. 2021YFB0300101, 62073333, 61902411, 62032023, 12002382, 11275269, 42104078), 173 Program of China (2020-JCJQ-ZD-029), Open Research Fund from State Key Laboratory of High Performance Computing of China (HPCL) (No. 202101-01), Guangdong Natural Science Foundation (2018B030312002), and the Program for Guangdong Introducing Innovative and Entrepreneurial Teams under Grant (No. 2016ZT06D211). Jose E. Roman is supported by the Spanish Agencia Estatal de Investigacion (AEI) under project SLEPc-DA (PID2019-107379RB-I00). On behalf of all authors, the corresponding author states that there is no conflict of interest.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Inertia-based spectrum slicing for symmetric quadratic eigenvalue problems

  • Campos, Carmen
  • Jose E. Roman|||0000-0003-1144-6772
[EN] In the quadratic eigenvalue problem (QEP) with all coefficient matrices symmetric, there can be complex eigenvalues. However, some applications need to compute real eigenvalues only. We propose a Lanczos-based method for computing all real eigenvalues contained in a given interval of large-scale symmetric QEPs. The method uses matrix inertias of the quadratic polynomial evaluated at different shift values. In this way, for hyperbolic problems, it is possible to make sure that all eigenvalues in the interval have been computed. We also discuss the general nonhyperbolic case. Our implementation is memory-efficient by representing the computed pseudo-Lanczos basis in a compact tensor product representation. We show results of computational experiments with a parallel implementation in the SLEPc library., Agencia Estatal de Investigacion, Grant/Award Number: TIN2016-75985-P
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A polynomial Jacobi-Davidson solver with support for non-monomial bases and deflation

  • Campos, Carmen
  • Jose E. Roman|||0000-0003-1144-6772
[EN] Large-scale polynomial eigenvalue problems can be solved by Krylov methods operating on an equivalent linear eigenproblem (linearization) of size d center dot n where d is the polynomial degree and n is the problem size, or by projection methods that keep the computation in the n-dimensional space. Jacobi-Davidson belongs to the latter class of methods, and, since it is a preconditioned eigensolver, it may be competitive in cases where explicitly computing a matrix factorization is exceedingly expensive. However, a fully fledged implementation of polynomial Jacobi-Davidson has to consider several issues, including deflation to compute more than one eigenpair, use of non-monomial bases for the case of large degree polynomials, and handling of complex eigenvalues when computing in real arithmetic. We discuss these aspects and present computational results of a parallel implementation in the SLEPc library., This work was supported by Agencia Estatal de Investigación (AEI) under Grant TIN2016-75985-P, which includes European Commission ERDF funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Non-linear eigenvalue problems with GetDP and SLEPc: Eigenmode computations of frequency-dispersive photonic open structures

  • Demesy, Guillaume
  • Nicolet, André
  • Gralak, Boris
  • Geuzaine, Christophe
  • Campos, Carmen
  • Jose E. Roman|||0000-0003-1144-6772
[EN] We present a framework to solve non-linear eigenvalue problems suitable for a Finite Element discretization. The implementation is based on the open-source finite element software GetDP and the open-source library SLEPc. As template examples, we propose and compare in detail different ways to address the numerical computation of the electromagnetic modes of frequency-dispersive objects. This is a non-linear eigenvalue problem involving a non-Hermitian operator. A classical finite element formulation is derived for five different solutions and solved using algorithms adapted to the large size of the resulting discrete problem. The proposed solutions are applied to the computation of the dispersion relation of a diffraction grating made of a Drude material. The important numerical consequences linked to the presence of sharp corners and sign-changing coefficients are carefully examined. For each method, the convergence of the eigenvalues with respect to the mesh refinement and the shape function order, as well as computation time and memory requirements are investigated. The open-source template model used to obtain the numerical results is provided. Details of the implementation of polynomial and rational eigenvalue problems in GetDP are given in the appendix. Program summary Program title: NonLinearEVP.pro CPC Library link to program files: http://dx.doi.org/10.17632/r57nxxtc62.1 Licensing provisions: GNU General Public License 3 Programming language: Gmsh (http://gmsh.info), GetDP (http://getdp.info) Nature of problem: Computing the eigenvalues and eigenvectors of electromagnetic wave problems involving frequency-dispersive materials. The resulting eigenvalue problem is non-linear and non-hermitian. Solution method: Finite element method coupled to efficient non-linear eigenvalue solvers: Relevant SLEPc solvers were interfaced to the Finite Element software GetDP. Several linearization schemes are benchmarked. (C) 2020 Elsevier B.V. All rights reserved., The work was partly supported by the French National Agency for Research (ANR) under the project "Resonance" (ANR-16-CE240013). The authors acknowledge the members of the project "Resonance'' for fruitful discussions. C. Campos and J. E. Roman were supported by the Spanish Agencia Estatal de Investigacion (AEI) under project SLEPc-HS (TIN2016-75985-P), which includes European Commission ERDF funds. C. Geuzaine was supported by ARC grant for Concerted Research Actions (ARC WAVES 15/19-03), financed by the Wallonia-Brussels Federation of Belgium. The authors thank Christian Engström from Ume¿ Universitet for helpful comments. Maxence Cassier from Institut Fresnel is acknowledged. Finally, the authors address special thanks to Anne-Sophie Bonnet Ben-Dhia and Camille Carvalho from INRIA (POEMS) for their motivating remarks and insights.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Thick-restarted joint Lanczos bidiagonalization for the GSVD

  • Alvarruiz Bermejo, Fernando|||0000-0001-5957-9561
  • Jose E. Roman|||0000-0003-1144-6772
  • Campos, Carmen
[EN] The computation of the partial generalized singular value decomposition (GSVD) of large-scale matrix pairs can be approached by means of iterative methods based on expanding subspaces, particularly Krylov subspaces. We consider the joint Lanczos bidiagonalization method, and analyze the feasibility of adapting the thick restart technique that is being used successfully in the context of other linear algebra problems. Numerical experiments illustrate the effectiveness of the proposed method. We also compare the new method with an alternative solution via equivalent eigenvalue problems, considering accuracy as well as computational performance. The analysis is done using a parallel implementation in the SLEPc library., This work was supported by Agencia Estatal de Investigacion (AEI), Spain under grant PID2019-107379RB-I00/AEI/10.13039/501100011033.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Acoustic modal analysis with heat release fluctuations using nonlinear eigensolvers

  • Hiremath, Varun
  • Jose E. Roman|||0000-0003-1144-6772
[EN] Closed combustion devices like gas turbines and rockets are prone to thermoacoustic instabilities. Design engineers in the industry need tools to accurately identify and remove instabilities early in the design cycle. Many different approaches have been developed by the researchers over the years. In this work we focus on the Helmholtz wave equation based solver which is found to be relatively fast and accurate for most applications. This solver has been a subject of study in many previous works. The Helmholtz wave equation in frequency space reduces to a nonlinear eigenvalue problem which needs to be solved to compute the acoustic modes. Most previous implementations of this solver have relied on linearized solvers and iterative methods which as shown in this work are not very efficient and sometimes inaccurate. In this work we make use of specialized algorithms implemented in SLEPc that are accurate and efficient for computing eigenvalues of nonlinear eigenvalue problems. We make use of the n-tau model to compute the reacting source terms in the Helmholtz equation and describe the steps involved in deriving the Helmholtz eigenvalue equation and obtaining its solution using the SLEPc library., Funding for open access charge: CRUE-Universitat Politècnica de Valéncia.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

An extension of the Cayley transform method for a parameterized generalized inverse eigenvalue problem

  • Dalvand, Zeynab
  • Hajarian, Masoud
  • Jose E. Roman|||0000-0003-1144-6772
[EN] Since recent studies have shown that the Cayley transform method can be an effective iterative method for solving the inverse eigenvalue problem, in this work, we consider using an extension of it for solving a type of parameterized generalized inverse eigenvalue problem and prove its locally quadratic convergence. This type of inverse eigenvalue problem, which includes multiplicative and additive inverse eigenvalue problems, appears in many applications. Also, we consider the case where the given eigenvalues are multiple. In this case, we describe a modified problem that is not overdetermined and discuss the extension of the Cayley transform method for this modified problem. Finally, to demonstrate the effectiveness of these algorithms, we present some numerical examples to show that the proposed methods are practical and efficient., The authors would like to express their heartfelt thanks to the editor and anonymous referees for their useful comments and constructive suggestions that substantially improved the quality and presentation of this article. This research was developed during a visit of Z.D. to Universitat Politecnica de Valencia. Z.D. would like to thank the hospitality shown by D. Sistemes Informatics i Computacio, Universitat Politecnica de Valencia. J.E.R. was partially supported by the Spanish Agencia Estatal de Investigacion (AEI) under grant TIN2016-75985-P, which includes European Commission ERDF funds. The authors thank Carmen Campos for useful comments on an initial draft of the article.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

KSPHPDDM and PCHPDDM: Extending PETSc with advanced Krylov methods and robust multilevel overlapping Schwarz preconditioners

  • Jolivet, Pierre
  • Zampini, Stefano
  • Jose E. Roman|||0000-0003-1144-6772
[EN] Contemporary applications in computational science and engineering often require the solution of linear systems which may be of different sizes, shapes, and structures. The goal of this paper is to explain how two libraries, PETSc and HPDDM, have been interfaced in order to offer end-users robust overlapping Schwarz preconditioners and advanced Krylov methods featuring recycling and the ability to deal with multiple right-hand sides. The flexibility of the implementation is showcased and explained with minimalist, easy-to-run, and reproducible examples, to ease the integration of these algorithms into more advanced frameworks. The examples provided cover applications from eigenanalysis, elasticity, combustion, and electromagnetism., Jose E. Roman was supported by the Spanish Agencia Estatal de Investigacion (AEI) under project SLEPc-DA (PID2019-107379RB-I00)
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Computation of scattering resonances in absorptive and dispersive media with applications to metal-dielectric nano-structures

  • Araujo C., Juan C.
  • Campos, Carmen
  • Engstrom, Christian
  • Jose E. Roman|||0000-0003-1144-6772
[EN] In this paper we consider scattering resonance computations in optics when the resonators consist of frequency dependent and lossy materials, such as metals at optical frequencies. The proposed computational approach combines a novel hp-FEM strategy, based on dispersion analysis for complex frequencies, with a fast implementation of the nonlinear eigenvalue solver NLEIGS. Numerical computations illustrate that the pre-asymptotic phase is significantly reduced compared to standard uniform h and p strategies. Moreover, the efficiency grows with the refractive index contrast, which makes the new strategy highly attractive for metal-dielectric structures. The hp-refinement strategy together with the efficient parallel code result in highly accurate approximations and short runtimes on multi processor platforms. (C) 2019 Elsevier Inc. All rights reserved., Juan C. Araujo and Christian Engstrom gratefully acknowledge the support of the Swedish Research Council under Grant No. 621-2012-3863. Carmen Campos and Jose E. Roman were supported by the Spanish Agencia Estatal de Investigacion (AEI) under project SLEPc-HS (TIN2016-75985-P), which includes European Commission ERDF funds. The supercomputer Tirant 3 used in some of the computational experiments belongs to Universitat de Valencia. Juan C. Araujo acknowledges Andree Falgin Hultgren, for his contributions on the meshing routine.