libMobility: a fast Stokesian dynamics library

ORAL

Abstract

We'll introduce a fast and modular GPU library, libMobility, that solves the hydrodynamic mobility problem for Stokesian particle suspensions in open, periodic, and confined geometries. In particular, we've implemented fast routines to apply mobility matrices and their principle square root, ensuring that the library is immediately available for use in Brownian dynamics applications. In addition to being a high-performance library, libMobility is designed to have a streamlined installation and be easy to integrate into existing codes. We present examples with fluctuating membranes, slender filaments, and active particles that demonstrates both the flexibility and the computational advantage of the library.

Presenters

  • Ryker Fish

    Colorado School of Mines

Authors

  • Ryker Fish

    Colorado School of Mines

  • Brennan Sprinkle

    Colorado School of Mines

  • Raúl Pérez Peláez

    Universitat Pompeu Fabra

  • Aleksandar Donev

    Courant Institute