GPU accelerated Entangled Star-Branched Polymer Rheology


Despite strong industrial interest in entangled polymer melts, there are very few models offering quantitative predictions of rheology. Such universal theory would allow commercial polymer blends to be modeled reliably and save billions of dollars on trial-and-error design. Discrete Slip-Link Models (DSM) developed by Schieber et al (1) offer such prediction power, relying on very few assumptions. Of particular interest are star-shaped polymers, whose relaxation times grow exponentially with size, which prevents examination of all but the shortest of them. The GPU-accelerated implementation of DSM, first described in (2) and further developed here, opens new possibilities in exploring star-shaped melts and blends with two orders of magnitude speedup over the previous CPU-only implementation.

San Jose, California