We see a speed-up from using scipy.fft, which uses the C++ FFT package pypocketfft, and we also implemented a few critical functions in compiled Rust, which provides further speed-ups for key operations.
We are also waiting for the upcoming OpenPIV C++ package, which is in preparation.
Use any of your preferred channels
uv pip install openpiv --upgrade
poetry install openpiv --upgrade
conda install openpiv --upgrade -c conda-forge
Please report any bugs, positive experiences, or ideas on Github issues of openpiv-python
Also, openpiv-python-examples was refurbished to use the new version, but you'll see quite regular openpiv-python notebooks; the code underneath is jumping from Rust to SciPy with the default ["auto"] mode, so you might not notice any changes
Best regards
Alex