topic You are welcome! in Intel® oneAPI Math Kernel Library

mkl fft3w choosing PRECISION SINGLE OR DOUBLE

Martin_Gumbau — Tue, 21 Oct 2014 08:56:15 GMT

Hi,

I have installed Intel Composer XE 14.0.2.144.

I can compile fft2w 'single' precision version and 'double' precision version adding a compilation parameter:

[PRECISION={MKL_DOUBLE|MKL_SINGLE}]

I do not known how to do the same in fftw3. The make command does not show me this option.

So I do not known how to compile my apps in single or double precision linking mkl fftw3 libraries.

Thanks in advance.

Hi, Martin

Evarist_F_Intel — Wed, 22 Oct 2014 04:26:04 GMT

Hi, Martin! FFTW3 interface are compiled for both precision at once. Just use fftw_* for double precision or fftwf_* for single precision in your application. Also, please note, that MKL has built-in interface for FFTW3. You can just link your fftw3-application with MKL instead of FFTW3 and it should work (using MKL).

Great Evarist,

Martin_Gumbau — Wed, 22 Oct 2014 09:08:32 GMT

Great Evarist,

I saw that you say in the code file names in .../mkl/interfaces/fftw3xc/wrapper but I wasn't sure about this.

Now all be clear.

Thanks a lot.

You are welcome!

Evarist_F_Intel — Wed, 22 Oct 2014 13:56:41 GMT

You are welcome!

SOLVED

Martin_Gumbau — Mon, 27 Oct 2014 13:18:54 GMT

SOLVED