GRISK: An Internet Based Search for K-Optimal Lattice Rules

This paper describe the implementation and underlying philosophie of a large scale distributed computation of K-optimal lattice rules. The computation is huge corresponding to the equivalent of 36 years computation on a single workstation. In this paper we describe our implementation, how we have built in fault tolerence and our strategy for administrating a world wide computation on computer we have no account on. We also provide a brief description of the problem, that of computing K-optimal lattice rules, and give some statistics to describe the extent of the search.