[Insight-developers] Dashboard issues to fix before final release

[Bradley Lowekamp]

I have std::transform a try:

http://review.source.kitware.com/#/c/11273/

Here is a table of timings of using CastImageFilter to convert from short to float:

gcc 4.4.7	std::copy	std::transform	std:copy speedup
[64, 64, 64]	0.00095768	0.002372	2.47654749
[128, 128, 128]	0.00751585	0.018858	2.509030915
[256, 256, 256]	0.06628470	0.163698	2.469619686
[32, 32]	0.00000294	0.000009	2.936786434
[64, 64]	0.00000306	0.000026	8.558496072
[128, 128]	0.00000704	0.000099	14.10766618
[256, 256]	0.00003077	0.000392	12.74427524


So... That is a bit slower with transform. I don't like it.

I have included the performance test in the patch too.

I also get the following compilation error with the patch:

ITK/Examples/Segmentation/CannySegmentationLevelSetImageFilter.cxx:376:   instantiated from here
ITK/Modules/Core/Common/include/itkImageAlgorithm.h:177: error: no matching function for call to ‘itk::CovariantVector<float, 2u>::CovariantVector(const itk::FixedArray<float, 2u>&)’



On May 14, 2013, at 12:03 PM, Bradley Lowekamp <blowekamp at mail.nih.gov> wrote:

> 
> We might be able to use std::transform in some cases... I'll check some timings...
> 
> Brad
> 
> On May 14, 2013, at 11:40 AM, Bill Lorensen <bill.lorensen at gmail.com> wrote:
> 
>> I don't think we should use cdash suppression for this one. Users will be annoyed, e.g. Slicer...
>> 
>> 
>> 
>> On Tue, May 14, 2013 at 11:38 AM, Bradley Lowekamp <blowekamp at mail.nih.gov> wrote:
>> 
>> As a last resort to getting ride of this warning, we could write an adaptor class. Provide an adaptor to a pointer to an iterator interface but on the dereference do the static_cast for the conversion. I don't know if the same performance benefits will still occur. And I don't really want to do this either, it'll be a lot more code.
>> 
>> Alternatively we could just add this warning to the CDash warning suppression file.
>> 
>> We could also just compile a function, and for just that file provide the command line ignore warning option.
>> 
>> The current code is correct, it's doing the right thing. We just don't want the warning...
>> 
>> Brad
>> 
>> On May 13, 2013, at 2:31 PM, Bradley Lowekamp <blowekamp at mail.nih.gov> wrote:
>> 
>> > Hello,
>> >
>> >
>> > On May 12, 2013, at 10:45 PM, Matt McCormick <matt.mccormick at kitware.com> wrote:
>> >>
>> >> Visual Studio possible loss of data warnings
>> >> https://issues.itk.org/jira/browse/ITK-3084
>> >
>> >
>> > I have tried a couple of things to deal with this warning:
>> >
>> > http://review.source.kitware.com/#/c/11222/
>> >
>> > I don't have another idea beside disabling that warning for all if the in itkWin32Header.h.
>> >
>> > Anyone have any suggestions?
>> >
>> > Brad
>> > _______________________________________________
>> > Powered by www.kitware.com
>> >
>> > Visit other Kitware open-source projects at
>> > http://www.kitware.com/opensource/opensource.html
>> >
>> > Kitware offers ITK Training Courses, for more information visit:
>> > http://kitware.com/products/protraining.php
>> >
>> > 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-developers
>> 
>> 
>> 
>> 
>> -- 
>> Unpaid intern in BillsBasement at noware dot com
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-developers/attachments/20130514/2a753e52/attachment.htm>