Fast Multipole Method (FMM) and Treecode are popular tree-based multipole algorithms with rigorous error estimates and wide applications in computing N-body particle interactions. FMM takes both near-field and far-field expansions, resulting in the revolutionary O(N) computations. Treecode takes only the far field expansion, receiving relatively less efficient O(N(logN)) computations while gaining saved memory and improved parallelization efficiency for high performance computing. This talk first compares both methods under the Cartesian expansion in terms of efficiency, memory use, and parallel performance. Following that, the idea of the multipole expansion is used on several biomolecular applications including pKa calculation, polarizable atomic multipole model, chromatin packing, and protein-protein interactions, all utilizing the distinguished feature of multipole methods in accelerating the calculation of electrostatic interactions.

Place: 316 meeting room, Envrionmental building, RUC