Gridap.jl: Efficient and accessible finite element software using the Julia programming language

26 Mai 2023
12:00
En liña
CITMAga

"Gridap.jl: Efficient and accessible finite element software using the Julia programming language"organizado por CITMAga. Será impartido por Francesc Verdugo (Computer Science Department, VU Amsterdam)

Data: venres 26 de maio

Hora: 12:00 h.

Duración: 50 min

Lugar: En liña (MS Teams)

Abstract:

The numerical solution of partial differential equations (PDEs) is among the most computationally intensive tasks in scientific computing. One often needs to deal with complex, nonlinear, multi-scale, and multi-physics problems that require to exploit the computational resources of large parallel computer clusters. The efficient solution of complex PDE-based models on such computing platforms often requires sophisticated software packages, which require a significant development effort and a steep learning curve for users. Julia is a new programming language specifically designed to achieve an enhanced trade-off between user experience and code performance. In particular, it has the potential to simplify the development, usage, and maintenance of scientific codes for the solution of PDEs and other mathematical problems. In this talk, we will discuss how we have leveraged the Julia programming language to build the Gridap.jl project, a finite element library exclusively written in Julia. These ideas could be potentially applied to other domains of interest.

Dr Francesc Verdugo is assistant professor at the High Performance Distributed Computing group at the Computer Science of Department of VU Amsterdam. His main research interests are the development of HPC simulation tools and the parallelization of numerical schemes, including different types of finite element (FE) methods and solvers for sparse systems of algebraic equations. F. Verdugo obtained his PhD at the Universitat Politècnica de Catalunya (UPC) and has been postdoc at the TU Munich and CIMNE before joining VU Amsterdam.