-
Notifications
You must be signed in to change notification settings - Fork 22
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add Test Dependency on ITKVkFFTBackend #159
Comments
@thewtex @dzenanz are there particular guidelines / design constraints that need to be considered for ITK remote modules depending on other ITK remote modules? In this case of |
When building ITK directly, enabling one remote module will automatically enable all other remote modules it depends on. When building externally, all the required dependencies must be enabled in the original ITK build. |
Thanks @dzenanz . It sounds like it would cause issues to have Is it reasonable to add |
That seems reasonable. We could add |
This can probably be done before 5.3 final. It might even be desirable for using 5.3 final in the ultrasound-related remote modules. |
Having a better understanding now of how FFT factory registration is intended to work, I'm not sure that we actually need @thewtex @dzenanz would it be reasonable for me to either remove this issue or amend so that ITKUltrasound may depend on ITKVkFFTBackend as a test dependency only? (for verifying speedup in an ultrasound context) |
Yes! |
We could avoid the module / C++ test dependency, and test in CI by installing the itk-vtkfftbackend package and running Python script tests. |
That would help reduce dependencies. One caveat on this issue is that we can't provide GPU-accelerated FFTs for curvilinear ultrasound images without one module having knowledge of the other. We may find ourselves needing to revisit this issue at some point in the future. |
ITKUltrasound
will depend onITKVkFFTBackend
for accelerated FFT computation.ITKVkFFTBackend
defines factory overrides for accelerated implementations of ITK FFT interface classes and should not be considered a direct dependency for ITKUltrasound. However, it would be valuable to test the user scenario where speedup on large ultrasound datasets can be accomplished with GPU-accelerated FFT.Similar to GPU testing in
ITKVkFFTBackend
, a GPU-based test should be added to demonstrate speedup when accelerated FFTs are registered.From ITKVkFFTBackend Issue 14.
The text was updated successfully, but these errors were encountered: