Related packages¶

Other recommended NUFFT libraries¶

NFFT3: well-supported and multi-featured C++ library using FFTW. Has MATLAB MEX interface. However, significantly slower and/or more memory-intensive than FINUFFT (see reference [FIN]). Has many more general abilities, eg, inverse NUFFT. We are working on this too.
CMCL NUFFT: NYU single-threaded Fortran library using self-contained FFT, fast Gaussian gridding kernel. Has MATLAB MEX interface. Much (up to 50x even for one thread) slower than FINUFFT, but very easy to compile.
cuFINUFFT: Our GPU version of FINUFFT, for single precision in 2D and 3D, type 1 and 2. Still under development by Melody Shih (NYU) and others. Often achieves speeds around 10x the CPU version.
MIRT Michigan Image Reconstruction Toolbox. Native MATLAB, single-threaded sparse mat-vec, prestores all kernel evaluations, thus is memory-intensive but surprisingly fast for a single-threaded implementation. However, slower than FINUFFT for all tolerances smaller than 0.1.
PyNUFFT Python code supporting CPU and GPU operation. Have not compared against FINUFFT yet.

Also see the summary of library performances in our paper [FIN] in the references.