Re: [ublas] FFT with ublas?

From: Peter Melchior (pmelchior_at_[hidden])
Date: 2007-03-19 11:59:40

Hello,

has anyone used FFTW3 with ublas? If yes, could you provide an example or show
me some pointers, please.

Best regards,

Peter Melchior

>>>>>> how typically do people utilize ublas with Fast Fourier Transform,
>>>>>> what FFT library do you use?
>>>>> We do this all the time using the Intel MKL libraries (also using the
>>>>> lapack bindings for the other bits).  I've done it before with the
>>>>> FFTW, but not recently.
>>>>
>>>> many thanks for your reply, I googled for "Intel MKL libraries" and
>>>> it turns out that this is a commercial library. In my circumstances
>>>> it is not acceptable. But http://www.fftw.org/ looks very promising.
>> I believe that the MKL is available free to non-commercial use, but you
>> are correct, the FFTW is probably more appropriate.  I last used FFTW2,
>> and it wasn't too bad to get it in with uBlas.  I know that FFTW3
>> changed a lot of things, I hope they weren't as short sighted to make
>> bindings difficult.
>>
>> Jeff
> 
> Here is the problem.  I think this fftw3 interface doesn't work well with my code:
> This is a quote from fftw3, about upgrading from fftw2:
> 
> Plans
> The major difference between FFTW 2 and FFTW 3 is in the planning/execution division
> of labor. In FFTW 2, plans were found for a given transform size and type, and then
> could be applied to any arrays and for any multiplicity/stride parameters. In FFTW 3,
> you specify the particular arrays, stride parameters, etcetera when creating the plan, and
> the plan is then executed for those arrays (unless the guru interface is used) and those
> parameters only. (FFTW 2 had “specific planner” routines that planned for a particular
> array and stride, but the plan could still be used for other arrays and strides.) That is,
> much of the information that was formerly specified at execution time is now specified at
> planning time.