[Insight-users] Re sampleImageFilter and multithreading.

[Luis Ibanez]

Hi Pete79,

Thanks for letting us know that you found the root of the problem.

On retrospective,
check for uninitialized variables may have helped to find this faster....


About your second question:

     We will need to know more about your processes....
     Are they running only the resampling of the image ?
     or do they have other components ?


I would suggest that you start by just running one of the
ITK Resampling examples with a large image and check
that the 8 cores are being used (they should).

You will find many examples of Resampling in the
directory:

           Insight/Examples/Filtering/

ResampleImageFilter2.cxx
ResampleImageFilter3.cxx
ResampleImageFilter4.cxx
ResampleImageFilter5.cxx
ResampleImageFilter6.cxx
ResampleImageFilter7.cxx
ResampleImageFilter8.cxx
ResampleImageFilter9.cxx
ResampleImageFilter.cxx
ResampleOrientedImageFilter.cxx
ResampleVolumesToBeIsotropic.cxx



    Please let us know what you observe
    when you run that experiment.


            Thanks


                   Luis


----------------------------------------------------------------------
On Sun, Oct 18, 2009 at 6:18 PM, Pete79 <pkoniusz at hotmail.com> wrote:
>
>
> Pete79 wrote:
>>
>> I'm running multiple threads which launch a subroutine which consists of
>> the following:
>>
>
> Dear all. Shame to admit, though the chief issue was caused by my own
> mistake. When setting up parameters of the affine transform, I set up only
> firs 4 out of 6 matrix values which correspond to 2x2 covariance matrix.
> Nonetheless, there are two more coefficient which correspond to shift. They
> ended up initialised with random values (possibly large ones) - this kept
> causing crashes and other odd behaviour. Now, I don't even limit number of
> threads by SetNumberOfThreads per resampler and it works fine.
>
> The only thing is I haven't got expected boost in processing time. When I
> run two separate processes of the code, it seems often utilise only few
> percent of the 8 cores I've got. At scarce peaks it gets 30%. I expected it
> to reach 80/90 percent easily, especially that two independent processes run
> bunch of threads simultaneously .... Any thoughts on this?
>
> Many thanks!
> --
> View this message in context: http://www.nabble.com/ResampleImageFilter-and-multithreading.-tp25942385p25951068.html
> Sent from the ITK - Users mailing list archive at Nabble.com.
>
> _____________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Please keep messages on-topic and check the ITK FAQ at: http://www.itk.org/Wiki/ITK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.itk.org/mailman/listinfo/insight-users
>