[Insight-users] Release 2.0.0 slower than 1.6.0 and 1.8.1-Cygwin1.5.13, GCC 3.3.3-3

[Frederic Perez]

Thank you very much. I have updated my source files with your committed 
files and now the ImageMomentsCalculator problem is over.

Unfortunately the related fix didn't solve the efficiency loss that I'm
experiencing in some flood-fill applications---I mentioned this in my 
first post. (I don't know if the fix solved the 
OtsuMultipleThresholdsCalculator problem posted by Karthik 
Krishnan---so far I haven't used that class.) As time permits I'll try to 
learn the reason why this happens. I'll let you know if I succeed.

Thanks again,

Frederic Perez

On Mon, 7 Mar 2005 17:13:44 -0500, Miller, James V (Research)
<millerjv at crd.ge.com> wrote:
> I checked in a change to itkImageBase.h and itkImage.h to inline the
> the computations in the Transform* methods in a manner that should
> still work for GCC 3.4
> 
> They are available in the CVS version of ITK.  I'll ask Luis to patch
> the 2.0 release.
> 
> Thanks to everyone who helped track down this performance issue.  With
> the fix, the timings for the test routine went from 2.0 seconds to 0.07
> seconds (ImageMomentsCalculator).
> 
> Jim

<<snipped>>