FastHartleyTransform

Documentation	Code Status	Guidelines

This package provides a Julia implementation of the Fast Hartley Transform (FHT). It supports both CPUs and NVIDIA CUDA GPUs by using following FFT (fast Fourier transform) kernels:

• CPU implementation uses FFTW.jl.
• NVIDIA GPU implementation uses CUDA.jl's CUDA.CUFFT module.

The implementation is built up on AbstractFastHartleyTransforms.jl, providing generic interfaces including fht(x), fht!(x), ifht(x), ifht!(x), plan_fht(x), and plan_ifht(x).

Enzyme differentiation and Reactant compilation are optional. See the documentation for usage, supported backends, and performance measurements.