Background: The monodomain equation serves as a key mathematical model in the field of cardiac electrophysiology, providing a description of the electrical activity occurring within cardiac tissue. However, numerical solution of this equation, especially on fine meshes, requires significant computational resources. Existing software tools such as openCARP and cbcbeat (based on the FEniCS project) offer platforms for conducting these simulations, but their operational speed can be a barrier to clinical use.
Aims: This research seeks to introduce an innovative open-source program that can quickly resolve monodomain equations. Our aim is to improve cardiac electrophysiology modeling by trying to surpass the speed and preserve the precision of established software tools such as cbcbeat and openCARP. The study analyses the efficiency of the new software compared to these current computational tools.
Methods: This newly developed program is written in C, specifically designed to optimize the numerical solution of monodomain equations. To assess its accuracy and performance, we replicate the Niederer benchmark simulations across the three software codes. The measured wall time of these simulations is broken into solver components and across different temporal and spatial resolutions. Each solver employs operator splitting methods to separate the ODE and PDE computation steps. Consistency is maintained by using equivalent numerical approaches for each software package.
Results: Preliminary performance evaluations indicate that openCARP exceeds cbcbeat by a factor of 8-9. Our solver shows competitive speed performance; it also exhibits robust computational stability and produces outcomes that align with those derived from globally accepted cardiac electrophysiology models.
Conclusions: The findings indicate that the new open-source software holds promise as a tool for facilitating real-time clinical decision-making in cardiac electrophysiology. To further enhance performance, future improvements will focus on expanding the software's capabilities, including the integration of GPU support.