作者: Hong‐Qiang Ding , Naoki Karasawa , William A Goddard III , None
DOI: 10.1063/1.463935
关键词:
摘要: The N2 computations implicit in the Coulomb and other long range interactions remain critical bottleneck atomic‐level simulations of structure dynamics large systems. We report here cell multipole method which scales linearly with N requires only modest memory. To demonstrate feasibility this approach, we systematic calculations on realistic polymer systems up to 1.2 million atoms a laboratory workstation. becomes faster than exact for 300 atoms, million‐atom polymer, it is 2377 times faster. treats class form qiqj/rijp, includes (p=1), London dispersion (p=6), or shielded (p=2) interactions. This well suited highly parallel vector computers.