Papers by Lucia Catabriga
Although the participation of women in Computing has decreased considerably in recent decades, th... more Although the participation of women in Computing has decreased considerably in recent decades, this topic has been increasingly discussed in Brazil and worldwide. Seeking to understand how this topic has been addressed in Computing, this article presents a systematic mapping of the literature on SBC publications in the last years. The mapping collected 128 publications, which were extracted, classified and analyzed according to the research questions, showing that few events publish on the subject, most publications refer to data analysis, most contexts involve motivation and empowerment of women. women in the area and that there is research being applied in practically all of Brazil. Resumo. Embora a participação das mulheres na computação tenha diminuído consideravelmente nasúltimas décadas, essa temática tem sido cada vez mais discutida no Brasil e no mundo. Buscando compreender como essa temática tem sido abordada na computação, este artigo apresenta um mapeamento sistemático da literatura sobre as publicações da SBC nosúltimos anos. O mapeamento coletou 128 publicações, que foram extraídas, classificadas e analisadas de acordo com as questões de pesquisa, evidenciando que poucos eventos publicam sobre o assunto, a maioria das publicações se refereà análise de dados, a maioria dos contextos envolve motivação e empoderamento das mulheres ná area e que há pesquisa sendo aplicada em praticamente todo o Brasil.
Revista Caribeña de Ciencias Sociales
In this work we present a nonlinear multiscale finite element method combined with local precondi... more In this work we present a nonlinear multiscale finite element method combined with local preconditioning for solving compressible Euler equations in conservative variables. The formulations are based on the strategy of separating scales, in which it is the core of the variational multiscale (finite element) methodology. The subgrid scale space is defined using bubble functions that vanish on the boundary of the elements, allowing to use a local Schur complement to define the resolved scale problem. The resulting numerical procedure allows the fine scales to depend on time. The formulation proposed added artificial viscosity isotropically in all scales of the discretization. Due to the fact that, density-based schemes suffer with undesirable effects of low speed flow including low convergence rate and loss of accuracy, local preconditioning is applied to the set of equations in the continuous case. We evaluate the multiscale formulation with local preconditioning in the low Mach numb...
Proceeding Series of the Brazilian Society of Computational and Applied Mathematics, Feb 20, 2020
In this work we present a nonlinear multiscale finite element method coupled with the Weiss-Smith... more In this work we present a nonlinear multiscale finite element method coupled with the Weiss-Smith/Choi-Merkle (WSCM) local preconditioner for solving steady compressible flows at low Mach numbers. The multiscale formulation is based on the strategy of separating scales, in which the subgrid scale space is spanned by bubble functions, allowing to use a static condensation procedure in the local matrix system to define the resolved scale problem. The resulting numerical formulation is completed by adding an artificial viscosity operator in all scales of the discretization. We evaluate the multiscale formulation coupled with the WSCM preconditioner comparing it with the non-preconditioned case. The numerical experiments show that this numerical methodology yields good results.
Proceeding Series of the Brazilian Society of Computational and Applied Mathematics, Apr 14, 2017
Resumo. Este trabalho apresenta um estudo numérico de soluções aproximadas para o problema de Hem... more Resumo. Este trabalho apresenta um estudo numérico de soluções aproximadas para o problema de Hemker utilizando os conceitos da Análise Isogeométrica (AIG) e a ferramenta GeoPDEs. O estudo avalia a influência da disposição das patches e dos knots na qualidade da solução aproximada, além de apresentar a influência da continuidade das funções base no tamanho e na esparsidade do sistema linear associado e no seu tempo de resolução.
Proceeding Series of the Brazilian Society of Computational and Applied Mathematics, Nov 18, 2015
Neste trabalho, utilizamos o algoritmo paralelo híbrido SPIKE como um precondicionador para um mé... more Neste trabalho, utilizamos o algoritmo paralelo híbrido SPIKE como um precondicionador para um método iterativo não-estacionário combinando as arquiteturas de memória distribuída e compartilhada MPI/OpenMP. A fim de obter um bom precondicionador resolvemos um conjunto de problemas combinatórios como reordenamento, particionamento, matching e o problema quadrático da mochila. Apresentamos os resultados avaliando o speedup e escalabilidade em sistemas resultantes de formulações de elementos finitos.
Proceeding Series of the Brazilian Society of Computational and Applied Mathematics, Apr 14, 2017
In this work we evaluate two multiscale methodologies to solve compressible flow problems, named,... more In this work we evaluate two multiscale methodologies to solve compressible flow problems, named, Dynamic Diffusion (DD) and Nonlinear Multiscale Viscosity (NMV), using the well know predictor-multicorrector time integration scheme. The subgrid scale space is defined using bubble functions whose degrees of freedom are locally eliminated in favor of the degrees of freedom that live on the resolved scales. The time integration schemes assume that the resolved coarse scale advances in time by second order approximation and the unresolved scale can advance by first and second order approximations. Performance and accuracy comparisons are conducted based on benchmark 2D problems.
Predictor-Multicorrector Schemes for the Multiscale Dynamic Diffusion Method to solve Compressible flow problems
Anais do ... Congresso Ibero-Latino-Americano de Métodos Computacionais em Engenharia, 2015
In this paper, we use the parallel hybrid SPIKE algorithm as a preconditioner for a nonstationary... more In this paper, we use the parallel hybrid SPIKE algorithm as a preconditioner for a nonstationary iterative method combining distributed and shared memory architectures, MPI and OpenMP. In order to obtain a good preconditioner we solve a set of combinatorial problems such as reorderings and graph partitioning. We present the results evaluating the influence of each strategy on the convergence and CPU time of the iterative solver. Resumo. Neste trabalho, utilizamos o algoritmo paralelo híbrido SPIKE como um precondicionador para um método iterativo não-estacionário combinando as arquiteturas de memória distribuída e compartilhada, MPI e OpenMP. A fim de obter um bom precondicionador, resolvemos um conjunto de problemas combinatórios como reordenamentos e particionamento de grafos. Apresentamos os resultados avaliando a influência de cada estratégia na convergência e tempo de CPU do método iterativo.
Local Preconditioning Techniques Coupled with a Variational Multiscale Method to Solve Compressible Steady Flows at Low Mach Numbers
Lecture Notes in Computer Science, 2019
In this work we evaluate a combination of the Weiss-Smith and Choi-Merkle local preconditioners c... more In this work we evaluate a combination of the Weiss-Smith and Choi-Merkle local preconditioners coupled with the density-based Nonlinear Multiscale Viscosity (NMV) finite element method for solving steady compressible flows at low Mach numbers. The multiscale formulation is based on the strategy of separating scales, in which the subgrid scale space is spanned by bubble functions, allowing to use a static condensation procedure in the local matrix system to define the resolved scale problem. Also, a residual-based nonlinear viscosity operator is added to the Galerkin formulation in order to obtain a stabilized formulation. As density-based methods do not work well in problems with Mach numbers tending to zero, resulting in a degradation of the solution accuracy, the resulting numerical method gathering those two approaches allows to solve compressible flows in the incompressible limit. We evaluate this methodology simulating a steady flow over the NACA 0012 airfoil under some regimes of inflow Mach numbers. The numerical result exhibits promising solutions to compressible flow problems in the incompressible limit.
A Nonlinear Subgrid Stabilization Parameter-Free Method to Solve Incompressible Navier-Stokes Equations at High Reynolds Numbers
Lecture Notes in Computer Science, 2019
In this work we evaluate a Nonlinear Subgrid Stabilization parameter-free method to solve time-in... more In this work we evaluate a Nonlinear Subgrid Stabilization parameter-free method to solve time-independent incompressible Navier-Stokes equations (NSGS-NS) at high Reynolds numbers, considering only the decomposition of the velocity field (not pressure) into coarse/resolved scales and fine/unresolved scales. In this formulation we use a dynamic damping factor which it is often essential for the nonlinear iterative process and for the reduction of the number of iterations. In order to reduce the computational costs typical of two-scale methods, the unresolved scale space is defined using bubble functions whose degrees of freedom are locally eliminated in favor of the degrees of freedom that live on the resolved scales. Accuracy comparisons with the streamline-upwind/Petrov-Galerkin (SUPG) formulation combined with the pressure stabilizing/Petrov-Galerkin (PSPG) are conducted based on 2D steady state benchmark problems with high Reynolds numbers, flow over a backward-facing step and lid-driven square cavity flow.
Convergence analysis of a new dynamic diffusion method
Computers & mathematics with applications, Sep 1, 2021
Trends in Applied and Computational Mathematics, Jan 10, 2018
The Reverse Cuthill-McKee (RCM) algorithm is a well-known heuristic for reordering sparse matrice... more The Reverse Cuthill-McKee (RCM) algorithm is a well-known heuristic for reordering sparse matrices. It is typically used to speed up the computation of sparse linear systems of equations. This paper describes two parallel approaches for the RCM algorithm as well as an optimized version of each one based on some proposed enhancements. The first one exploits a strategy for reducing lazy threads, while the second one makes use of a static bucket array as the main data structure and suppress some steps performed by the original algorithm. These related changes led to outstanding reordering time results and significant bandwidth reductions. The performance of two algorithms is compared with the respective implementation made available by Boost library. The OpenMP framework is used for supporting the parallelism and both versions of the algorithm are tested with large sparse and structural symmetric matrices.
Thermo-Economic Simulation of the Thermal Insulated Equipment Used in Steam Injection Wells by Finite Volumes Method
An Alternative Approach of the SPIKE Preconditioner for Finite Element Analysis
This paper describes an alternative approach to using the SPIKE preconditioner in the context of ... more This paper describes an alternative approach to using the SPIKE preconditioner in the context of finite element analysis. Commonly, SPIKE operations evaluate finite element matrices in sequential form with the resulting matrices distributed in parallel partitions. The goal is to obtain these matrices literally in parallel, avoiding excessive communication at the beginning of the process. This study offers a comparison of two SPIKE versions. One version considers a domain decomposition of the mesh as a preprocess for the finite element calculations. The other version considers a domain decomposition of the resulting linear system. With numerical experiments, this paper shows that the first version demonstrates advantages related to CPU time and memory usage.
The objective of this work is to present, analyze and discuss the distribution of genders in Comp... more The objective of this work is to present, analyze and discuss the distribution of genders in Computer Science and Computer Engineering courses at the Federal University of Espírito Santo (UFES), with the aim of investigating the current low female presence. Studies in the world have verified that there has been a great reversal in genders in computing, previously occupied by women creators of various technologies and programming languages. This reality is repeated at UFES, being measured in this work through data, provided by the university, of students entering courses from 1990 to 2022. The quantitative analysis considered, for each student, the year of admission, age, gender, course and form of evasion. In addition to the descriptive analyses, Pearson's chi-square statistical tests were performed in order to measure the influence of gender on course completion. Resumo. O objetivo deste trabalhoé apresentar, analisar e discutir a distribuição dos gêneros 1 nos cursos de Ciência da Computação e Engenharia de Computação da Universidade Federal do Espírito Santo (UFES), com o intuito de investigar a baixa presença feminina atual. Estudos no mundo verificaram que houve uma grande inversão nos gêneros na computação, antes ocupada por mulheres criadoras de várias tecnologias e linguagens de programação. Essa realidade se repete na UFES, sendo mensurada nesse trabalho através dos dados, cedidos pela universidade, dos estudantes ingressantes nos cursos de 1990 a 2022. Na análise quantitativa foram considerados, para cada estudante, o ano de ingresso, a idade, o sexo, o curso e a forma de evasão. Além das análises descritivas, testes estatísticos de qui-quadrado de Pearson foram executados com o intuito de mensurar a influência do gênero na conclusão dos cursos.
O presente trabalho estuda o desempenho da paralelizac ¸ão do método dos elementos finitos utiliz... more O presente trabalho estuda o desempenho da paralelizac ¸ão do método dos elementos finitos utilizando estratégia de decomposic ¸ão de domínio com estruturas de blocos orientados da matriz de discretizac ¸ão resultante e três formatos de armazenamento de matrizes esparsas. O sistema linear de equac ¸ões proveniente da formulac ¸ão do método dos elementos finitos é resolvido através do método iterativo não-estacionário GM-RES. Os esquemas de armazenamento empregam vers ões paralelas da estratégia elemento por elemento, aresta por aresta e do tradicional formato de linhas esparsas comprimidas. A implementac ¸ão é desenvolvida para arquiteturas de memória distribuída, particularmente para clusters de estac ¸ões de trabalho, e a troca de mensagens entre os processadores é efetuada através da biblioteca MPI.
Parallel algorithms capable to use efficiently thousands of multi-core processors is the trend in... more Parallel algorithms capable to use efficiently thousands of multi-core processors is the trend in High Performance Computing. To achieve a high scalability, hybrid solvers are suitable candidates since they can combine the robustness of direct methods and the low computational cost of iterative methods. The parallel hybrid SPIKE algorithm has been used as a preconditioner on a non-stationary iterative solver combining distributed and shared memory architectures MPI/OpenMP. In this work we apply several combinatorial problems [2] such as reorderings, partitionings and matchings to obtain a good preconditioner. We present results of speedup and scalability as well as comparisons of different approaches for the combinatorial algorithms in a set of large benchmark matrices from CISE [3].
The streamline-upwind/Petrov-Galerkin (SUPG) formulation of compressible flows based on conservat... more The streamline-upwind/Petrov-Galerkin (SUPG) formulation of compressible flows based on conservation variables, supplemented with shock-capturing, has been successfully used over a quarter of a century. In this paper, for inviscid compressible flows, the YZ shock-capturing parameter, which was developed recently and is based on conservation variables only, is compared with an earlier parameter derived based on the entropy variables. Our studies include comparing, in the context of these two versions of the SUPG formulation, computational efficiency of the element-and edge-based data structures in iterative computation of compressible flows. Tests include 1D, 2D, and 3D examples.
A Trade-off Analysis of the Parallel Hybrid SPIKE Preconditioner in a Unique Multi-core Computer
Lecture Notes in Computer Science, 2017
In this paper we apply the parallel hybrid SPIKE algorithm as a preconditioner for a nonstationar... more In this paper we apply the parallel hybrid SPIKE algorithm as a preconditioner for a nonstationary iterative method to solve large sparse linear systems. In order to obtain a good preconditioner, we employ several strategies solving combinatorial problems such as matching, reordering, partitioning, and quadratic knapsack. Our SPIKE implementation combines MPI and OpenMP paradigms in a unique multi-core computer. The computational experiments show the influence of each strategy evaluating the number of iterations and CPU time of the iterative solver in a set of large systems from miscellaneous application areas. The experiments suggest that the SPIKE preconditioner is very advantageous when a suitable set of parameters is chosen. The choice of the number of MPI ranks and OpenMP threads is not an easy task, because the SPIKE algorithm can increase the number of iterations when the number of MPI ranks grows. Moreover, the increase in the number of threads does not ensure a better performance.
Studies of a Hybrid and Multiscale Tumor Growth Model via Isogeometric Analysis using PetIGA
This work presents experiments considering a hybrid and multiscale model of avascular and vascula... more This work presents experiments considering a hybrid and multiscale model of avascular and vascular tumor growth. The model uses phase-field equations to describe the tumor and vascular growth, and reaction-diffusion equations for the distribution of nutrients and angiogenic factors. We use the high-performance framework PetIGA based on Isogeometric Analysis to implement the model. Our numerical experiments show that the model is able to represent the capillary network formation and its ability to represent specific tumor growth configurations.
Uploads
Papers by Lucia Catabriga