Dresden 2014 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
CPP: Fachverband Chemische Physik und Polymerphysik
CPP 58: Polymer Dynamics II
CPP 58.6: Vortrag
Freitag, 4. April 2014, 11:00–11:15, ZEU 114
A Highly Parallelizable Monte Carlo Method For Simulating Polymeric Systems — •Christoph Jentzsch1, Ron Dockhorn1, Marco Werner1, and Jens-Uwe Sommer1,2 — 1Leibniz-Institut für Polymerforschung Dresden, Germany — 2TU-Dresden, Intitut für theoretische Pyhsik, Germany
We show a new variant of the Bond Fluctuation Model (BFM), a coarse grained algorithm to simulate polymer systems. Our model is based on a bcc lattice instead of a simple cubic lattice used for the original BFM. This leads to a modifed set of bond vectors, as well as larger moves. The algorithm allows for a very efficient parallelization which is based in subdividing the bcc lattice in two disjunct sublattices. By implementing our algorithm on a consumer graphic card, we gain an upspeed of about two orders of magnitude as compared to the CPU- implementation. We compare our model with the original BFM for dense polymer melts with respect to static and dynamic properties. We conclude that this model is applicable to simulate all quantities and polymeric systems as the original BFM does, but with a significant upspeed.