If you are comfortable with building PDL yourself,
I would suggest you install version 2 of FFTW
and make sure the PDL::FFTW module is built
in addition to modifying NFACTORS.  That way
you are covered either way.  (I don't know the
FFT code so it is possible that the small value
of NFACTORS may be required for some
reason...  [ You may have to build FFTW2
by hand as a lot of package managers have
stopped supporting it.  Another reason for
wanting PDL::FFTW3. ]

Good luck,
Chris

On Fri, Feb 15, 2013 at 10:38 AM, Frank Boers <[email protected]> wrote:
> Hi Chris,
> I m using debian wheezy 64bit and  PDL v2.4.11.
> PDL is installed manual via make on a network harddisk and local on my
> notebook.
> With the newer versions it is really easy ti install!!
> Several PCs and our cluster are using the same PDL version.
> I will update to the new PDL version soon and try to change the NFACTORS.
>
> best regards
> Frank
>
>
>
> On 15.02.2013 16:06, Chris Marshall wrote:
>>
>> Hi Frank-
>>
>> See the PDL web page at http://pdl.perl.org/?page=install
>> for ways to get PDL.  You don't say what your platform is
>> but PDL should build out of the box if you have development
>> environment set up.
>>
>> If you have any problems or questions regarding the PDL
>> install/build guides, please follow up on the perldl list.
>> I'm sure someone can help and we'd like to keepp the
>> instructions as up-to-date and correct as possible.
>>
>> --Chris
>>
>> On Fri, Feb 15, 2013 at 7:40 AM, Frank Boers<[email protected]>
>> wrote:
>>>
>>> Dear Chris,
>>> thank you, that explains a lot and my question is now 100% answered.
>>> I m not sure how to go ahead may rebuild PDL with a different value for
>>> FACTORS. or waiting for the ffftw3 implementation.
>>>
>>> best regards
>>> Frank
>>>
>>>
>>>
>>> On 14.02.2013 17:09, Chris Marshall wrote:
>>>>
>>>>
>>>> Looking at the source, the maximum number
>>>> of factors is 11 and your number is 2**23 or
>>>> 4**11.5 which is why the fail.  You can either:
>>>> - rebuild PDL with a different value for NFACTORS
>>>> - pick a number that has fewer factors
>>>> - install FFTW2 and rebuild PDL to use it
>>>>
>>>> For example, if I use a size of 5**10 then the
>>>> FFT completes with the existing code.  I don't
>>>> know about speed tradeoffs.
>>>>
>>>> Cheers,
>>>> Chris
>>>>
>>>> On Thu, Feb 14, 2013 at 10:22 AM, Frank Boers<[email protected]>
>>>> wrote:
>>>>>
>>>>>
>>>>> Dear Craig and Chris,
>>>>> thank you for suggestions.
>>>>> I checked the following with a linux system 8Gb in total with padded
>>>>> numbers. The data I have to process are less than 4194304 elements, so
>>>>> I
>>>>> m
>>>>> fine.
>>>>>
>>>>>
>>>>> use PDL::FFT
>>>>>
>>>>> $d=random(4194304)
>>>>> p $d->info('%M')  # 32.00MB
>>>>>
>>>>> realfft($d) # =>   OK
>>>>>
>>>>> #--- next step
>>>>> $d=random(8388608)
>>>>> p $d->info('%M') # 64.00MB
>>>>>
>>>>> realfft($d)
>>>>>
>>>>> Error: fftradix() - exceeded number of factors
>>>>> Error: fftradix() - insufficient memory.
>>>>>
>>>>> best regards
>>>>> Frank
>>>>>
>>>>> ...snip...
>
>
> --
> Frank Boers
> Institute of Neuroscience and Medicine - 4
> Medical Imaging Physics
> Forschungszentrum Juelich GmbH
> 52425 Juelich
> phone: +49 - (0)2461-61-6005
> fax  : +49 - (0)2461-61-2820
> email: [email protected]
> http://www.fz-juelich.de/inm/inm-4/DE/Forschung/MEG-Physik/_node.html
>
>
>
> ------------------------------------------------------------------------------------------------
> ------------------------------------------------------------------------------------------------
> Forschungszentrum Juelich GmbH
> 52425 Juelich
> Sitz der Gesellschaft: Juelich
> Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
> Vorsitzender des Aufsichtsrats: MinDir Dr. Karl Eugen Huthmacher
> Geschaeftsfuehrung: Prof. Dr. Achim Bachem (Vorsitzender),
> Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt,
> Prof. Dr. Sebastian M. Schmidt
> ------------------------------------------------------------------------------------------------
> ------------------------------------------------------------------------------------------------

_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl

Reply via email to