Theory and Algorithms for Parallel Computation

Bridging models such as BSP and LogP are playing an increasingly important role in scalable parallel computing. They provide a convenient architectureindependent framework for the design, analysis, implementation and comparison of parallel algorithms and data structures. The Theory and Algorithms sessions at EuroPar '98 will cover a number of important issues concerning models of parallel computation. The topics covered will include: The BSP model and its relationship to LogP, BSP algorithms for PC clusters, distributed shared memory models, computational complexity of parallel algorithms, BSP algorithms for discrete event simulation, cost modelling of shared data types, systolic algorithms, cache optimisation, optimal scheduling of task graphs.