Digital twins in cardiac electrophysiology offer a powerful framework for simulating patient-specific cardiac behavior, enabling the personalization of diagnosis and therapy. These applications require numerical solvers that can efficiently handle biophysically detailed models in anatomically accurate geometries. MonoAlg3D is an open-source software package developed to meet these requirements by solving the three-dimensional anisotropic monodomain equation using the Finite Volume Method (FVM) on structured meshes.
The codebase is primarily written in C, with CUDA-based GPU acceleration used for solving both the reaction and diffusion components of the model. Reaction and diffusion terms are decoupled using operator splitting, and the user may supply any cellular model compatible with MonoAlg3D's interface. The software includes two default solvers for ordinary differential equations: an explicit Euler method and an adaptive time-step variant of the explicit Euler method, allowing users to balance performance and accuracy according to the characteristics of the chosen ionic model.
MonoAlg3D is designed with a modular architecture, enabling users to extend or replace core components through well-defined interfaces. Custom functions can be provided for a variety of tasks, including mesh loading, ECG computation, and output in different mesh file formats. This flexibility facilitates the integration of MonoAlg3D into broader simulation pipelines and allows its adaptation to diverse research needs.
The software has been employed in high-throughput simulation studies involving hundreds of digital twins on GPU clusters, supporting applications such as parameter sensitivity analysis, model personalization, and conduction system calibration. Through its computational efficiency, extensibility, and open-source availability, MonoAlg3D provides a robust foundation for advancing cardiac digital twin technologies.