MKL FFT

The speed-boosted variants of NumPy’s FFT operations are accessible in the numpy.fft package, and the accelerate.mkl.fftpack package. The following functions in these packages are accelerated using MKL:

Function Description
fft(a, n=None, axis=-1) 1-dimensional forward transform
ifft(a, n=None, axis=-1) 1-dimensional inverse transform
rfft(a, n=None, axis=-1) 1-dimensional forward transform of purely real data
irfft(a, n=None, axis=-1) 1-dimensional inverse transform of purely real data
hfft(a, n=None, axis=-1) Hermite transform
ihfft(a, n=None, axis=-1) Inverse Hermite transform
fftn(a, s=None, axes=None) N-dimensional forward transform
ifftn(a, s=None, axes=None) N-dimensional inverse transform
fft2(a, s=None, axes=(-2, -1)) 2-dimensional forward transform
ifft2(a, s=None, axes=(-2, -1)) 2-dimensional inverse transform
rfftn(a, s=None, axes=None) N-dimensional forward transform of purely real data
rfft2(a, s=None, axes=(-2, -1)) 2-dimensional forward transform of purely real data
irfftn(a, s=None, axes=None) N-dimensional inverse transform of purely real data
irfft2(a, s=None, axes=(-2, -1)) 2-dimensional inverse transform of purely real data

For further information on these functions, please refer to the Numpy documentation: numpy.fft.