[gmx-users] install single & double precision versions of fftw-3.0.1

[Jian Zou]

David wrote:
> You need only the flag --enable-float for single
> fftw.h is indentical for single and double
> Just compile with and without the flag.
> Check the lib directory afterwards.

Hi David,

I checked the dir fftw3/lib, and found some files. They may be sorted as
libfftw3.*    double
libfftw3_threads.*    double & mpi
libfftw3f.*    single
libfftw3f_threads.*    single & mpi
Am I right?

Thanks again for your reply.


Regards,

Jian Zou