[Insight-users] Computation time

[Joshua Cates]

In your CMake configuration under CMAKE_BUILD_TYPE type "Release".  This
will cause CMake to use the CMAKE_CXX_FLAGS_RELEASE (-O2 or -O3 on most
platforms)  when compiling.  Another option is to leave the
CMAKE_BUILD_TYPE empty and modify CMAKE_CXX_FLAGS directly to -O3.

You will need to do a clean build for the changes to have any effect.  I
would suggest having both an ITK Release and and ITK Debug build tree for 
doing development.

Josh.

______________________________
 Josh Cates			
 Scientific Computing and Imaging Institute
 University of Utah
 Email: cates at sci.utah.edu
 Phone: (801) 587-7697
 URL:   http://www.sci.utah.edu/~cates


On Wed, 18 Feb 2004, Radhika Sivaramakrishna wrote:

> Hi Josh,
> How do I make the build optimized?
> I am using pixel-wise iteration.
> Radhika
> 
> ----- Original Message ----- 
> From: "Joshua Cates" <cates at sci.utah.edu>
> To: "Radhika Sivaramakrishna" <radshashi at earthlink.net>
> Cc: <insight-users at itk.org>; "Luis Ibanez" <luis.ibanez at kitware.com>
> Sent: Wednesday, February 18, 2004 9:51 AM
> Subject: Re: [Insight-users] Computation time
> 
> 
> > Hi Radhika,
> >
> > Is your ITK build optimized (i.e. -03)?  ITK depends in some cases on
> > compiler optimizations for efficiency.
> >
> > Another factor may be the type of the iterator that you are using.
> > Are you doing neighborhood or pixel-wise iteration?
> >
> > Josh.
> >
> > ______________________________
> >  Josh Cates
> >  Scientific Computing and Imaging Institute
> >  University of Utah
> >  Email: cates at sci.utah.edu
> >  URL:   http://www.sci.utah.edu/~cates
> >
> >
> > On Wed, 18 Feb 2004, Radhika Sivaramakrishna wrote:
> >
> > > Hi Luis,
> > > I implemented a program to segment T1-weighted 3D brain images. I
> > > implemented this first in IDL and then also reimplemented this in ITK. I
> > > confirmed that the results I get are the same with both implementations.
> I
> > > find that the IDL program runs in a fraction of time compared to the ITK
> > > one. With IDL, it is just less than a minute whereas with IDL it is a
> good
> > > 10-15 minutes for each image. Is this something that you would expect or
> am
> > > I making some mistake? I would have thought that because IDL is higher
> > > level, that it would need more time.
> > >
> > > I have used all the standard filters in ITK, and connected them in a
> > > pipeline [typical filters [apart from Read, Write etc] are
> > > BinaryThresholdImageFilter, ConnectedComponentImageFilter,
> > > RelabelComponentImageFilter etc]. I have also had to use iterators to do
> > > some processing.
> > > What could be the computationally intensive part, the pipeline of
> filters or
> > > the use of iterators?
> > >
> > > Thanks
> > > Radhika
> > >
> > >
> > > _______________________________________________
> > > Insight-users mailing list
> > > Insight-users at itk.org
> > > http://www.itk.org/mailman/listinfo/insight-users
> > >
> >
> >
> 
> 
>