Edlyn TESKE (Waterloo)
Abstract: The Pollard kangaroo method is used to compute discrete logarithms in arbitray cyclic groups if the discrete log is known to lie in a certain interval. In contrast to the baby step-giant step method, it is very space efficient and can be parallelized with linear speed up.
In our talk, we discuss the parallelized kangaroo method and show how to apply it to the infrastructure in real quadratic function fields. We announce the computation of a 29-digit regulator of a random real quadratic function field of odd characteristic and of genus 3, which is the largest such computation done so far.