-
Notifications
You must be signed in to change notification settings - Fork 0
Add Tpetra-based MueLu tutorial driver and fix hands-on.py (without rebase) #30
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 Tpetra-based MueLu tutorial driver and fix hands-on.py (without rebase) #30
Conversation
Signed-off-by: Philip Oesterle-Pekrun <[email protected]>
Signed-off-by: Philip Oesterle-Pekrun <[email protected]>
…ver.cpp, and fix plotting in hands-on.py Signed-off-by: Philip Oesterle-Pekrun <[email protected]>
…xample for the tutorial and replacement for ScalingTest.cpp Signed-off-by: Philip Oesterle-Pekrun <[email protected]>
… TutorialDriver.cpp Signed-off-by: Philip Oesterle-Pekrun <[email protected]>
Signed-off-by: Philip Oesterle-Pekrun <[email protected]>
|
CDash for AT1 results [Only accessible from Sandia networks] |
|
I am honestly super confused why clang-format is failing. The format_patch has a bunch of global files, not any of the files I edited in my commits. Even the only MueLu file, |
|
@PhilipOesterlePekrun I am fine with the current changes. @jhux2, how should we proceed? Can we merge into your branch? |
Signed-off-by: Philip Oesterle-Pekrun <[email protected]>
|
@PhilipOesterlePekrun This looks good. It looks like |
Signed-off-by: Philip Oesterle-Pekrun <[email protected]>
|
@jhux2 Sorry about that, I forgot to remove |
|
Thanks! |
|
@mayrmt @PhilipOesterlePekrun Any objections if I squash commits so that we can merge to Trilinos? |
|
@jhux2 No objections from my side. |
|
@jhux2 feel free to go ahead! |
|
Great. Just to clarify, in the branch https://github.com/jhux2/Trilinos/tree/new-muelu-read-the-docs, there are ~80 older commits by @mayrmt and @GrahamBenHarper that are missing a sign-off in their commit messages (because they predate that requirement). Is it ok if I squash those commits? I would need to sign off myself. I would indicate in the commit that Matthias and Graham were the original authors. |
|
@jhux2 I am fine with that. |
Equivalent to #29 but without rebasing. That is, some ctests would fail when this is merged into jhux2 due to a not-yet-present change in Galeri at this point. But, it will succeed when merged into the main Trilinos repository's develop branch (so, the tests in trilinos#14357 should still succeed when this is merged into jhux2).
@mayrmt
@jhux2
@trilinos/MueLu
Motivation
This change serves mainly to migrate the executables/drivers used in
hands-on.pyfrom the Epetra to Tpetra backend. This was most easily done by reducing the number of .cpp files--which were doing a lot of the same things--into a single generalTutorialDriver.cpp. This is now is the driver for Laplace2D, Recirc2D, the challenge problems, and any future problem types we might want to introduce.Details
The challenge problems are not exactly the same as before because
https://trilinos.org/wordpress/wp-content/uploads/2015/07/MueLu_tutorial_challenges.tar.gzis currently not accessible, but they still serve the same purpose (inefficient reference xml, which the user should try to optimize).ScalingTest.cppandScalingTestParamList.cppwere removed for now because there is no explicit mention of scaling in the tutorial, and much of the functionality was unrelated and convoluted (especially if a tutorial user wants to look into the files directly). For theincludeliterals in the "C++ Interface" subsection of the "Using MueLu in User Applications" section, the shorterTutorial_cppInterface.cppis instead used, which concisely and cleanly shows the C++ interface without using an xml file.The
CMakeLists.txtoftest/tutorial/was also streamlined to test theTutorialDriver.cppwith xml files which are explicitly mentioned in the tutorial.The
TutorialDriver.cppnow writes the outputsexample<PID>.txt, which are to be used for gnuplots inhands-on.py, as in the original Epetra version oflaplace2d.cpp. Some small fixes inhands-on.pyensure that this functionality works properly now.Related Issues
Part of trilinos#14357
Stakeholder Feedback
Testing