[Insight-developers] ITK test failures in Debug only

[Luis Ibanez]

Hi Mathieu,

Thanks a lot for trying with ITK 3.2.

It is rather bad news....

I interpret that as:

     "the error has been there for a long time"

On the bright side, you are not getting
these errors with your VTK build, so we
could trace this by comparing the vtkMultithreader
with the itkMultithreader.

Just to double check:

    Is your VTK build using also Debug ?
    Is it submitted to the VTK Dashboard ?


I will suggest that we identify which one of
the current failing test is the simplest case
and then we track from it.

The fact that this happens only in a Debug build
still leaves open the option that it might be
an unitialized variable that happens to be in
the Multithreader.

Is it possible to run Valgrind in one of these tests ?

Uninitialized Memory Reads (UMR) could be potential suspects....



     Luis



------------------------
Mathieu Coursolle wrote:
> Hi,
> 
> Just to let you know, I did a clean build of ITK 3.2 and
> the behavior is the same for Debug. There are still some test
> failures of the same type.
> 
> Mathieu
> 
> 
>>On 9/13/07 11:56 AM, Luis Ibanez said:
>>
>>
>>>Fixing this one is going to be *a lot* of fun.
>>
>>Yup! :(  Thank god it's happening in Debug.
>>
>>
>>>I'm wondering if the failures in "mini2"
>>>http://www.itk.org/Testing/Sites/mini2/Darwin-Xcode2.2.1-dbg/
>>>20070913-0100-Nightly/Test.html
>>>are due to the same phenomenon.
>>
>>How many CPUs/cores does that machine have?  If you ssh to it, you can
>>run the 'system_profiler' command to determine this.
>>
>>
>>>   Is this a something that started failing recently in your build ?
>>>   or do we have reasons to believe that it has never worked ?
>>
>>We've never had an ITK Debug dashboard before.  We created it for MINC2
>>testing and chose Debug only because our other one is Release.
>>
>>However, Frank's dashboard (4 core Mac Pro) has been failing for
>>months.  Oldest I can find is:
>><http://www.itk.org/Testing/Sites/SINTEF.LabMAC.OSX/
>>ITKcvs_CMake246_gcc401/20070309-0100-Nightly/Test.html>
>>
>>>   Does your same machine builds VTK ?
>>>   if so, do you see any similar errors in VTK ?
>>
>>Yes; No.
>>
>>
>>>   The reason why I'm asking is because the ITK Multithreader
>>>   descends from the VTK one.
>>>
>>>   The started with the same genes, but have evolved in different
>>>   ecological niches.
>>>
>>>   A gvmdiff comparisson of both brings interesting "__APPLE__"
>>>   genes that only the ITK MultiThreader has acquired.
>>>   E.g. Lines 123-134.
>>
>>I've actually changed that one Apple-only part... It is the same in VTK
>>and ITK.  It's pretty harmless, it just determines how many cores/CPUs
>>there are and uses that as the default # of threads to create.
>>
>>I'm no threading expert, but I did give itkMultiTheader.cxx a quick
>>lookover.  It is weak on error checking.  All the pthread_... APIs
>>return error codes, but the code almost never checks them.
>>
>>
>>>If it is not too much to ask, I'll be interested in seeing if
>>>a build of ITK 3.2 in your system brings up these same errors.
>>>If they do not appear, at least we will have framed the problem
>>>in a 6-months window.
>>
>>Mathieu is trying now...
>>
>>-- 
>>____________________________________________________________
>>Sean McBride, B. Eng                 sean at rogue-research.com
>>Rogue Research                        www.rogue-research.com 
>>Mac Software Developer              Montréal, Québec, Canada
>>
>>
>>
> 
> 
> 
>