Monte Carlo simulations of quantum systems on massively parallel supercomputers

Abstract

An efficient Monte Carlo algorithm for two‐dimensional quantum spins or boson systems was implemented on two massively parallel distributed memory supercomputers: Intel Delta and iPSC/860. The algorithm uses a multispin coding technique which allows significant data compactification and efficient vectorization of Monte Carlo updates. Two data decompositions, corresponding naturally to different Monte Carlo updating processes and observable measurements, are devised such that only nearest neighbor communications are needed within a given decomposition. The algorithm dynamically switches between the two decompositions using a simple pipe along the one‐dimensional ring of processing nodes of the parallel computer. In this two‐decomposition scheme, the entire algorithm uses only two communication routines (exchange and broadcast); this simplification makes the code easily portable to other computers. On a 128‐node Intel Delta, this algorithm updates 183 million spins per second (21 on CM‐2 and 6.2 on Cray YMP). A systematic performance analysis shows a better than 90% efficiency in the parallel implementation. Several new physics results on the magnetic properties of the high temperature superconductors obtained in the simulations will be discussed.