High performance solver of the multidimensional generalized nonlinear Schrodinger equation with coupled fields

• Tipo: Artigo em Livro de Atas de Conferência Internacional
• Data: 2019

Autores

Ferreira, TD (Autor) Outra; Silva, NA (Autor) Outra; Guerreiro, A (Autor) FCUP

Ata de Conferência Internacional

Título: FOURTH INTERNATIONAL CONFERENCE ON APPLICATIONS OF OPTICS AND PHOTONICS

4th International Conference on Applications of Optics and Photonics (AOP)

Univ Lisbon, Lisbon, PORTUGAL, MAY 31-JUN 04, 2019

Indexação

ISI Web of Knowledge - 0 Citações

Scopus - 4 Citações

Outras Informações

ID Authenticus: P-00R-8E7

DOI: 10.1117/12.2527305

Abstract (EN): We report on the development of numerical module for the HiLight simulation platform based on GPGPU supercomputing to solve a system of coupled fields governed by the Generalized Nonlinear Schrodinger Equation with local and/or nonlocal nonlinearities. This models plays an important role in describing a plethora of different phenomena in various areas of physics. In optics, this model was initially used to describe the propagation of light through local and/or nonlocal systems under the paraxial approximation, but more recently it has been extensively used as a support model to develop optical analogues. However, establishing the relation between the original system and the analogue, as well as, between their model and the actual experimental setup is not an easy task. First and foremost because in most cases the governing equations are nonintegrable, preventing from obtaining analytical solutions and hindering the optimization of the experiments. Alternatively, despite numerical methods not providing exact solutions, they allow to test different experimental scenarios and provide a better insight to what to expect in an actual experiment, while giving access to all the variables of the optical system being simulated. However, the numerical solution of a system of N-coupled Schrodinger fields in systems with two or three spatial dimensions requires massive computation resources, and must employ advanced supercomputing and parallelization techniques, such as GPGPU. This paper focuses on the numerical aspects behind this challenge, describing the structure of our simulation module, its performance and the tests performed.

Idioma: Inglês

Tipo (Avaliação Docente): Científica

Nº de páginas: 7