Parallel algorithms for simulating interacting carriers in nanocommunication

Abstract

Diffusion-based molecular communication (DMC) with interacting carrier molecules allows exploring a variety of new communication paradigms. Simulating these interactions on a network scale is however hard from an analytic point of view, and continuous detection of which couples of particles may interact is a computationally expensive task. This work proposes different algorithms that allow detecting which particles are interacting. Hierarchical grid based approaches are explored and designed to maximally exploit the parallelism offered by multiple-graphics processing units (multi-GPU) systems. An implementation of the proposed algorithm is compared to some of our previous work in this domain and to a state-of-the-art collision detection library and nanosimulator. Performance tests indicate a speedup of up to a factor 100 compared to the collision detection algorithm implemented in the nanosimulator, and a factor 77 relative to the collision detection library.