mkl_fft started as a part of Intel (R) Distribution for Python* optimizations to NumPy, and is now being released
as a stand-alone package. It can be installed into conda environment using
conda install -c intel mkl_fft
Since MKL FFT supports performing discrete Fourier transforms over non-contiguously laid out arrays, MKL can be directly used on any well-behaved floating point array with no internal overlaps for both in-place and not in-place transforms of arrays in single and double floating point precision.
This eliminates the need to copy input arrayy contiguously into an intermediate buffer.
mkl_fft directly supports N-dimensional Fourier transforms.
More details can be found in SciPy 2017 conference proceedings: https://github.com/scipy-conference/scipy_proceedings/tree/2017/papers/oleksandr_pavlyk
It implements the following functions:
fft(x, n=None, axis=-1, overwrite_x=False)
ifft(x, n=None, axis=-1, overwrite_x=False)
fft2(x, shape=None, axes=(-2,-1), overwrite_x=False)
ifft2(x, shape=None, axes=(-2,-1), overwrite_x=False)
fftn(x, n=None, axes=None, overwrite_x=False)
ifftn(x, n=None, axes=None, overwrite_x=False)
rfft(x, n=None, axis=-1, overwrite_x=False) - real 1D Fourier transform, like scipy.fftpack.rfft
rfft_numpy(x, n=None, axis=-1) - real 1D Fourier transform, like numpy.fft.rfft
rfft2_numpy(x, s=None, axes=(-2,-1)) - real 2D Fourier transform, like numpy.fft.rfft2
rfftn_numpy(x, s=None, axes=None) - real 2D Fourier transform, like numpy.fft.rfftn
... and similar irfft* functions.
The package also provides mkl_fft._numpy_fft and mkl_fft._scipy_fft interfaces which provide drop-in replacements for equivalent functions in NumPy and SciPy respectively.
To build mkl_fft from sources on Linux:
- install a recent version of MKL, if necessary;
- execute
source /path/to/mklroot/bin/mklvars.sh intel64; - execute
pip install .