Date of creation: 20151009

Type of project:

Internship or final degree project

Candidate profile:

Student of Computer Science or of Physics/Mathematics with a good background in computing and some knowledge of HPC.

Description:

The Earth Sciences Department of the Barcelona Supercomputing Center (BSC-ES) offers an internship/final degree project on “Data Assimilation code for HPC platforms”. Data assimilation is a statistical technique which combines model simulations of a real system with measurements. Its main application is for forecast initialisation in atmospheric and oceanic sciences, however, it can solve a variety of state estimation problems in geophysics and engineering. This project will give the participant the unique opportunity to run complex numerical model simulations within an HPC environment, and to develop skills for parallel programming.

The project consists of the following three main tasks to perform on the BSC-ES data assimilation code:

• Multi-processor parallelisation

This will require implementing MPI code changes for efficiently running the code on multi-processors.

• Multi-scale implementation

This will require the implementation in FORTRAN language of a flexible model domain with a solution to the problem of the domain’s boundary conditions.

• Performance improvement

This will require the analysis of the efficiency of the code performance using the state-of-the-art tools developed by the Computer Sciences Department of BSC.