Balance Parallelization Of Dissipative Particle Dynamics

As a widely concerned technique, parallel computing and cloudcomputing which is based on parallel computing have become the preferredmethod of dealing with mass data processing and complicated computingproblems. More and more applications are moved from local to cloud. Thedevelopment of the parallel computing combines the computing and memoryresources and improves the computing capabilities. Dissipative ParticleDynamics (DPD) is a method of molecular simulation based on mesoscopicsystem. And now it is widely used in polymer chemistry, biology and otherrelated areas. But the problems simulated by DPD are usually large-scale, andit takes a very long time to simulation an ordinary problem. In order to reducethe computing time of DPD simulation, a balance parallelization of dissipativeparticle dynamics (BPDPD) method is presented in this paper, whichcombines the parallel computing and DPD together. The experiments onLinux cluster show that the BPDPD can not only effectively reduce thecomputing time of DPD, but also make the load on the cluster more balanceand reduce the data amount transmitted between computing nodes.In this paper, firstly, the basic method of DPD is briefly introduced.And then the algorithm of BPDPD is presented in detail. After that, theaggregation of anchored proteins is introduced as an application example of DPD simulation to test the performance of our program. At last, the BPDPDprogram is tested on a Linux cluster, and an experiment report is shown in thispaper.