[Insight-users] Why each thread has a separate copy of the transform in itkImageToImageMetric

[Mengda Wu]

Hi all,

   I am looking at the code in itkImageToImageMetric.txx (latest version in
git). I have a question why each thread needs to have a
seperate copy of the transform (variable m_ThreaderTransform[i]). Why not
use a single one as in thread 0 (variable m_Transform)? Should
all transforms be the same at all times? Or, they can be changed by the
optimizer in each thread and thus we need synchronize
wtih SynchronizeTransforms().

  Any clarification or document of the multi-threading strategy of ITK's
metric and optimizer is greatly welcomed.

Thanks,
Mengda
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20100909/59173737/attachment.htm>