Cardiac electrophysiology numerical models using symmetric multiprocessing (SMP)

Multi-dimensional electrophysiological models have been introduced to investigate electrical propagation in tissue level, based on cell-dynamics models. The models include a set of non-linear differential equations which describe the dynamics of cell and tissue excitation. However, as models evolve, it is inevitable that proper and powerful tools need to be introduced in order to reproduce the detailed and thus computationally intensive simulations. To build such tools, several computational methodologies need to be adopted regarding efficiency and reliability. On the other hand improvements apply to the hardware too. State of the art computers, even personal computers, tend to make use of multiple core Central Processing Units. Unfortunately the aforementioned methodologies follow sequential logic, resulting to low efficiency of the working platform. In this work we present the performance bottleneck in symmetric multiprocessing (SMP) for simulations of propagation phenomena in cardiac tissue electrophysiological models. We demonstrate the scalability and efficacy of the different methodologies used in the discretisation scheme and message passing in SMP.