Developer Reference for Intel® oneAPI Math Kernel Library for C

ID 766684
Date 11/07/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

Multi-threaded FFTW

This section discusses multi-threaded FFTW wrappers only. MPI FFTW wrappers, available only with Intel® oneAPI Math Kernel Library (oneMKL) for the Linux* and Windows* operating systems, are described in a separate section.

Unlike the original FFTW interface, every computational function in the FFTW2 interface to Intel® oneAPI Math Kernel Library (oneMKL) provides multithreaded computation by default, with the maximum number of threads permitted in FFT functions (see "Techniques to Set the Number of Threads" in Intel® oneAPI Math Kernel Library (oneMKL) Developer Guide). To limit the number of threads, call the threaded FFTW computational functions:

void fftw_threads(int nthreads, fftw_plan plan, int howmany, fftw_complex *in, int istride, int idist, fftw_complex *out, int ostride, int odist);

void fftw_threads_one(int nthreads, rfftwnd_plan plan, fftw_complex *in, fftw_complex *out);

...

void rfftwnd_threads_real_to_complex( int nthreads, rfftwnd_plan plan, int howmany, fftw_real *in, int istride, int idist, fftw_complex *out, int ostride, int odist);

Compared to its non-threaded counterpart, every threaded computational function has threads_ as the second part of its name and additional first parameter nthreads. Set the nthreads parameter to the thread limit to ensure that the computation requires at most that number of threads.

Product and Performance Information

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.

Notice revision #20201201