Photonic and RF simulation and inverse design in just few lines of code! Fully featured, fully differentiable, GPU accelerated FDTD engine and geometry optimizer. Experimental release: expect occasional critters..
Untitled.1.mp4
Features
- Automatic differentiation (AD) compatible for adjoint optimization
- Optional GPU acceleration
- Length scale controlled geometry optimizer
- Nonlinear and anisotropic materials
- Tensor subpixel smoothing for accuracy
- Adaptive graded grid and Float16 support for speed
- PML, periodic, PEC, PMC boundaries
- Modal sources, plane waves, Gaussian beams
- Modal monitors, DFT fields
- Concise mathematical syntax for update equations
Application specific Python API UI and gdsfactory integration for simulation and inverse design
- Photonic integrated circuits (PIC) - available now
- Metasurface optics and thin films - please request
- RF microstrip and patch atennas - please request
General purpose backend API is in Julia and is not documented (unless there's interest). Look at the beauty of Maxwell's equations in our update loop! Other FDTD packages make it so ugly :(
Follow us on our new Linkedin page for package updates and bug fixes! Feel free to raise an issue. We respond daily. Star us if you like this repo :)
Paul Shen, pxshen@alumni.stanford.edu