maxwell-equations

fdtd3d is an open source 1D, 2D, 3D FDTD electromagnetics solver with MPI, OpenMP and CUDA support for x64, ARM, ARM64, RISC-V, PowerPC, Wasm architectures

PyPEEC - 3D Quasi-Magnetostatic FFT/PEEC Solver

Photonic crystals in Julia 🦚

Full-wave 3D electromagnetic field simulation in Python

Time Domain Maxwell Solver

An nanophotonics solver for inverse design of metamaterials

USEL is an open-source CPU-based scientific computing framework for computational physics and quantum research. It provides modular numerical tools, reusable components, and developer-friendly APIs for building scientific applications. Designed for efficiency, portability, and accessibility, it runs on standard hardware without GPU requirements.

A self-contained 2D finite-element electrostatics solver for simulating real capacitor geometries — parallel plates, coaxial cables, and arbitrary shapes built from simple primitives — rather than relying on closed-form formulas that only exist for a handful of idealized geometries. Pure NumPy / SciPy / Matplotlib.

An FDFD electromagnetic mode solver for rasterized waveguide cross sections made to be a standard plugin for FDTD engines.

PyTorch-based electromagnetic FDTD simulator in pure Python for 1D, 2D, and 3D domains, with CPU and multi-GPU CUDA execution

2D FDTD solver of Maxwell's equations

My lectures for the University of Liverpool PHYS201 (Electromagnetism I) module

Official implementation of STAR-PINN: a Physics-Informed Neural Network (PINN) for solving the 2D time-domain magnetic diffusion (Maxwell) equation in nonlinear materials.

Solving the Eddy Current problem with a multiplicative Schwarz smoothed geometric multigrid preconditioned GMRES solver in deal.II.

Lecture Notes in Electromagnetics II for Engineering Physics master students at Uppsala University (2022-2026)

AI-assisted electromagnetic cavity eigenmode simulator with deterministic Maxwell/Nedelec FEM, natural-language orchestration, C++17/Python, Gmsh and PETSc/SLEPc

Finite-difference time-domain (FDTD) method for modeling computational electrodynamics

Physics-informed neural network solving Maxwell's equations in ICP reactors. Hard BC ansatz, transfer learning across geometries, autograd sensitivity maps. 1700× faster than FEM. Built with NVIDIA Modulus + PyTorch.

Explore computational physics in Threejs VR

FEniCS programs for solving Maxwell's equations in some special cases