[Insight-developers] ITK 4.4 release candidate cycles about to begin

[Bradley Lowekamp]

Sean,

As I recall Hans had something to do with adding those lengthy test some time ago. They were to simulate real world data and registration.

It appears that ITK_USE_BRAINWEB_DATA needs to be enable for these tests.

Should be still be supporting these tests? If so we are doing a poor job of it.

Sean, If you run the tests without a timeout does is finish?

Brad


On May 8, 2013, at 3:25 PM, Sean McBride <sean at rogue-research.com> wrote:

> On Wed, 8 May 2013 14:36:08 -0400, Bradley Lowekamp said:
> 
>> http://open.cdash.org/viewTest.php?onlyfailed&buildid=2898774
>> 
>> These "DeformableRegistration8Tests" that are timing out. I don't think
>> any one else is running them. I think that they are enabled when memory
>> reaches a certain size.
> 
> Ah.  I do have ITK_COMPUTER_MEMORY_SIZE:STRING=28
> 
>> Perhaps you are running into resource limitation
>> with these tests? Running too many taking at the same time where each is
>> trying to use all the cores or too much memory? Or there is another bug
>> with the Transforms that is causing an infinite loop.
> 
> I do have it set up to run tests in parallel (takes too long otherwise).
> 
> Just tried now running only DeformableRegistration8Test5 with the machine idle, and it still times out:
> 
> 2441: Test timeout computed to be: 1500
> 2441: 
> 2441: Starting Registration 2441: 0   -1.48608   0
> 2441: 0   -1.48851   0
> 2441: 0   -1.48851   0.000128687
> 2441: 1   -1.48898   0.000128687
> 2441: 1   -1.48898   0.00011346
> 2441: 2   -1.49183   0.00011346
> 2441: 2   -1.49183   0.000145161
> 2441: 3   -1.49452   0.000145161
> 2441: 3   -1.49452   0.000136456
> 2441: 4   -1.49961   0.000136456
> 2441: 4   -1.49961   0.000201535
> 2441: 5   -1.51107   0.000201535
> 2441: 5   -1.51107   0.000146507
> 2441: 6   -1.53322   0.000146507
> 2441: 6   -1.53322   0.000196887
> 2441: 7   -1.54577   0.000196887
> 2441: 7   -1.54577   0.000153671
> 2441: 8   -1.55584   0.000153671
> 2441: 8   -1.55584   0.000217068
> 2441: 9   -1.56199   0.000217068
> 2441: 9   -1.56199   0.000156194
> 2441: 10   -1.56632   0.000156194
> 2441: 10   -1.56632   0.000186157
> 2441: 11   -1.57364   0.000186157
> 2441: 11   -1.57364   0.000167512
> 2441: 12   -1.58199   0.000167512
> 2441: 12   -1.58199   0.000161963
> 2441: 13   -1.58848   0.000161963
> 2441: 13   -1.58848   0.000129904
> 2441: 14   -1.59635   0.000129904
> 2441: 14   -1.59635   0.000187408
> 2441: 15   -1.60149   0.000187408
> 2441: 15   -1.60149   0.000205094
> 2441: 16   -1.60776   0.000205094
> 2441: 16   -1.60776   0.000199981
> 2441: 17   -1.61606   0.000199981
> 2441: 17   -1.61606   0.000192347
> 2441: 18   -1.62312   0.000192347
> 2441: 18   -1.62312   0.000192474
> 2441: 19   -1.62995   0.000192474
> 2441: 19   -1.62995   0.000161226
> 2441: 20   -1.63554   0.000161226
> 2441: 20   -1.63554   0.000128248
> 2441: 21   -1.63997   0.000128248
> 2441: 21   -1.63997   0.000140027
> 2441: 22   -1.64383   0.000140027
> 2441: 22   -1.64383   0.000153305
> 2441: 23   -1.64788   0.000153305
> 2441: 23   -1.64788   0.000135753
> 2441: 24   -1.65238   0.000135753
> 2441: 24   -1.65238   0.000191106
> 2441: 25   -1.65605   0.000191106
> 2441: 25   -1.65605   0.000168328
> 2441: 26   -1.65954   0.000168328
> 2441: 26   -1.65954   0.000131302
> 2441: 27   -1.66331   0.000131302
> 2441: 27   -1.66331   0.000147456
> 2441: 28   -1.66537   0.000147456
> 2441: 28   -1.66537   0.000174976
> 2441: 29   -1.66722   0.000174976
> 2441: Optimizer stop condition = LBFGSBOptimizer: Too many iterations. Iterations = 30
> 2441:           Probe Tag    Starts    Stops             Time (s)
> 2441:         Registration           1            1           1193.68
> 2441:           Probe Tag    Starts    Stops           Memory (kB)
> 2441:         Registration           1            1            117884
> 1/8 Test #2441: DeformableRegistration8Test5 .....***Timeout 1500.11 sec
> 
> The test pegs all cores, but doesn't use all that much memory that I noticed.
> 
> Perhaps this test is just doing too much?
> 
> Cheers,
> 
> -- 
> ____________________________________________________________
> Sean McBride, B. Eng                 sean at rogue-research.com
> Rogue Research                        www.rogue-research.com 
> Mac Software Developer              Montréal, Québec, Canada
> 
>