Weekly Update, Release 25.04

[EZoni]

Weekly update to latest AMReX, pyAMReX, PICSAR:

./Tools/Release/updateAMReX.py
./Tools/Release/updatepyAMReX.py
./Tools/Release/updatePICSAR.py

New WarpX release 25.04:

./Tools/Release/newVersion.sh

[EZoni]

I suggest that we wait for AMReX-Codes/amrex#4407 to be merged and then rerun the update here to include that bug fix, which is needed in other open PRs, e.g., #5768.

[EZoni]

@WeiqunZhang @atmyers

Are you aware of any recent changes in AMReX that could explain the fact that the checksums of the following 3D tests seem to need an update?

The following tests FAILED:
	273 - test_3d_particle_absorption.checksum (Failed)
	277 - test_3d_particle_scrape.checksum (Failed)
	281 - test_3d_particle_scrape_picmi.checksum (Failed)
	429 - test_3d_eb_picmi.checksum (Failed)
Errors while running CTest

I am thinking about changes (e.g., bug fixes) since the March release (we skipped a few updates), such as, for example,

• Fixed EB data for anisotropic grids: boundary centroid, normal vector (backward compatible) AMReX-Codes/amrex#4357
• Bug fix for index_assert in Array4 AMReX-Codes/amrex#4400
• Fix a bug in 1d Curl Curl Solver AMReX-Codes/amrex#4407

[WeiqunZhang]

Yes the EB change may cause roundoff errors. The other two are not bugs that change the solution. One is like instead of checking n < something, where n is 0, 1, 2, ..., it only checkes n = 0. The other is a harmless bug that could trigger fpe.

[EZoni]

Thanks for the update, @WeiqunZhang.

The relative errors seem larger than roundoff in some cases. But I guess the magnitude of those may depend on the resolution of the test and other factors as well.

I will push a commit to fetch the latest commit before AMReX-Codes/amrex#4357 in order to verify that the checksums changes seen here are indeed due to that bug fix. If that is the case, I will push the back the update and reset the checksums.

[EZoni]

With 67c23ab, fetching the latest commit before AMReX-Codes/amrex#4357, all checksums pass:
https://dev.azure.com/BLAST-WarpX/WarpX/_build/results?buildId=292&view=results.

Now testing ab8d4d8, fetching the merge commit of AMReX-Codes/amrex#4357...
https://dev.azure.com/BLAST-WarpX/WarpX/_build/results?buildId=294&view=results.

The following tests FAILED:
	273 - test_3d_particle_absorption.checksum (Failed)
	277 - test_3d_particle_scrape.checksum (Failed)
	281 - test_3d_particle_scrape_picmi.checksum (Failed)
	429 - test_3d_eb_picmi.checksum (Failed)
Errors while running CTest

The checksums changes seem to be caused indeed by AMReX-Codes/amrex#4357.

[EZoni]

I tested both the commit before AMReX-Codes/amrex#4357 was merged and the commit where AMReX-Codes/amrex#4357 was merged and I can confirm that the checksums changes come from this PR.

I pushed a commit to reset the checksums. This should be ready for another review once the CI checks have completed.