Feature Request: GPU Acceleration Support for Diffraction Calculations

[HeeneFZ]

First, thank you for developing and maintaining this useful library! I was wondering if you’ve considered adding GPU computation support (e.g., CUDA/PyTorch/TensorFlow) to the diffraction calculation functions for potential performance improvements?

Many numerical operations in wave optics (e.g., FFTs, large matrix operations) could benefit significantly from GPU parallelization. This might be especially helpful for large-scale simulations or iterative optimizations.

Would this be feasible/aligned with the project’s goals? I’d be happy to help test or contribute if needed.

Thanks for your time!

[optbrea]

Thank you very much for your commentary. I would like to improve the computation as you say, but I am developing the package by myself and I am slow. This year I will have an student that will work in GPU computation for diffractio algorithms. Any suggestion is welcomed.

Thanks.
Luis MIguel