Sunday, June 24, 2018
  [ login ]
events
Coffee Talk - Parallel octrees using shared and distributed memory scheme

Wednesday, June 15th, 2016. Time: 15h.

Place: OCZ Conference Room, C1 Building, UPC Campus Nord, Barcelona.

ABSTRACT

Nowadays numerical simulations in sciences and engineering are an important part in selection and design of new products, buildings, etc. This kind of simulations calculate an approximate solution about a problem over a domain which is determined by the user.

Most of numerical simulations usually needs a domain to be discretized into sub-domains. In Finite Element Method (FEM) the discretization result are elements which can have different shapes, sizes and number of nodes.

In this work is presented the use of octrees to generate meshes suited for FEM using parallel computing with shared (OpenMP) and distributed (MPI) memory schemes.

SPEAKER CV

Jorge López RuizIndustrial Engineer by the UAA. He is currently pursuing a master's degree in computing at CIMAT. Member of the group of optimization and modeling using numerical methods. His thesis topic is the meshing using octrees in parallel.