FFTW 3.x in Octave ( Was: Re: [Fwd: Re: rfftw slower than fftw for "bad" size arrays])

[Dmitri A. Sergatskov]

John W. Eaton wrote:
> On 29-Jan-2004, Dmitri A. Sergatskov <address@hidden> wrote:
>
> | > Shouldn't we be using the latest version, if possible?
> | > 
> | > I do remember the fftw malloc problem from some earlier attempt to use
> | > fftw.  If it is possible to avoid that, then we probably should.  I
> | 
> | It may or may not be related to http://fftw.org/faq/section3.html#leaks
> | 
> | <<<<
> |    Question 3.15. FFTW seems to have a memory leak.
> |    After you create a plan, FFTW caches the information required to quickly
> |    recreate the plan. (See Q3.8 `Can I save FFTW's plans?') It also maintains
> |    a small amount of other persistent memory. You can deallocate all of FFTW's
> |    internally allocated memory, if you wish, by calling fftw_cleanup(), as
> |    documented in the manual.
> |  >>>>
>
> I was thinking there was someone who reported a bug related to the
> 16-byte alignment requirement and who told me that I should switch to
> using fftw_malloc for all of Octave, but I said that wasn't possible.
>
> jwe

Well, according to
http://www.fftw.org/fftw3_doc/SIMD-alignment-and-fftw_malloc.html#SIMD%20alignment%20and%20fftw_malloc

<<<<
  You are not _required_ to use fftw_malloc. You can allocate your data in any 
way that you like,
  from malloc to new (in C++) to a static array declaration. If the array 
happens not to be
  properly aligned, FFTW will not use the SIMD extensions.
>>>>

Since fftw probably not the only package which would benefit from SIMD, I 
assume g++ people
has some pressure to make "new" to do proper alignement...
But except performance issue there is no requirement to do it everywhere.

Sincerely,

Dmitri.