Speaker: Filip Jeremic

Abstract:
Lattice basis reduction has been successfully used for many problems in integer programming, cryptography, number theory, and information theory. In this presentation, we discuss a parallel version of the lattice basis reduction algorithm called the Jacobi Method. The Jacobi Method is very captivating as it is inherently parallel. We take advantage of this by utilizing the graphics processing unit (GPU) to capitalize on the algorithm's parallel nature. This presentation will cover an introduction to GPU aided computation, an explanation of why the Jacobi Method can be parallelized, and finally we will discuss the run time statistics of this parallel version of the Jacobi Method.