"An Odysseus internship affords undergraduate students from around the world with an interest in Computer Science the opportunity to undertake exciting and fun research in a distributed yet cooperative environment."
Radio propagation models are used to describe the physical radio channel when planning wireless systems. Recently developed propagation models, such as those based on ray-tracing, offer high accuracy but can be slow to run on typical desktop PCs. This burden increases even further when one is considering for example, an ad hoc network, where one must model the radio link between many pairs of independent mobile users. One solution is to use a parallel computing resource, running the simulation for each channel on a separate processor. This project involves modifying existing C++ ray-tracing code such that it can run on the DCU Scientific Computing and Complex Systems (Sci-Sym) cluster. The student will be working in a team with postgraduate and post-doctoral researchers on an existing project.
The project is concerned with the user using sensed information (about their location and the prevailing channel conditions at that point) to make informed decisions about their choice of network. Consequently, this work is of direct relevance to the theme.
The student will produce parallelised C++ code which produces a demonstrable, quantified scaling speed up when run on the DCU Sci-Sym cluster.
The student will need a PC and remote access to the Sci-Sym cluster.