To learn parallel programing with FFTs using intel MKL

[Angel_Vergara]

hi everyone

Im working with FFTs routines using Intel MKL and I need to develop an application in parallel. Who couldtell me where I can to find documentation about to learn parallel programing withFFTs?


thanks

Angel

[TimP]

Such questions, if not covered in the MKL docs, would usually be asked on the MKL forum, at least if they pertain to parallelism within an FFT. Of course, if you have multiple independent FFT analyses of similar size, it may be more effective to make parallel tasks or loops outside the MKL calls, which makes the parallelism more generic. Then there is nothing specific to FFT, except that you may choose to invoke mkl_sequential.

[Angel_Vergara]

Hi Tim!

I forget to mention that I'm working with fortran 90 on Linux, OpenSuse11.1. Ive checked the MKL docs, but the examples are oriented to C++. You know where can to find examples in fortran?

effectively I have multiple independent FFT analyses, but I would like to reduce the process time becausein the lastcomputationthe time was increased considereable...

thanks

Angel

Quoting - tim18

Such questions, if not covered in the MKL docs, would usually be asked on the MKL forum, at least if they pertain to parallelism within an FFT. Of course, if you have multiple independent FFT analyses of similar size, it may be more effective to make parallel tasks or loops outside the MKL calls, which makes the parallelism more generic. Then there is nothing specific to FFT, except that you may choose to invoke mkl_sequential.

[mahmoudgalal1985]

Quoting - tim18

Such questions, if not covered in the MKL docs, would usually be asked on the MKL forum, at least if they pertain to parallelism within an FFT. Of course, if you have multiple independent FFT analyses of similar size, it may be more effective to make parallel tasks or loops outside the MKL calls, which makes the parallelism more generic. Then there is nothing specific to FFT, except that you may choose to invoke mkl_sequential.

Thank you

[barragan_villanueva_]

Hi,

Pleaselook atcorresponding KB article(MKL forum):
http://software.intel.com/en-us/articles/different-parallelization-techniques-and-intel-mkl-fft/

-- Victor