AMReX/PICSAR: Weekly Update

[ax3l]

Weekly update to latest AMReX.
Weekly update to latest PICSAR (no changes).

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

[ax3l]

@WeiqunZhang @atmyers this week's update of AMReX changes WarpX checksums. Do you have time to investigate this?

• 2D: background_mcc_dp_psp (position)
• 3D: pml_psatd_dive_divb_cleaning (rho)
• SP: reduced_diags_single_precision (momentum)

[ax3l]

@atmyers @PhilMiller looking at the changes, I think you know this best:
AMReX-Codes/amrex@9aa23c2...826cd37

Likely in this PR:

• Add roundoff_lo corresponding to roundoff_hi for domains that don't start at 0 AMReX-Codes/amrex#2950

[WeiqunZhang]

The failure of 3D: pml_psatd_dive_divb_cleaning (rho) was due to the change in amrex::Geometry::outsideRoundOffDomain.

[ax3l]

2D: background_mcc_dp_psp (position) - @peterscherpelz:

That seems fine to me, I don't have any reason a small change like that would matter. Thanks for checking 

SP: reduced_diags_single_precision (momentum) - @RemiLehe:

I think that we can also reset the test reduced_diags_single_precision:
the errors are consistent with potentially having a slightly different number of particles in the box, due to the change in the condition that removes particles.

3D: pml_psatd_dive_divb_cleaning (rho) - @EZoni:

That test has three different lasers, polarized in three different directions, and I think it was added to show that using div cleaning in the PML (together with div cleaning in the valid domain) helps mitigate some instability that was otherwise growing at the domain-PML interface. This is described in more detail in PR [#1600](https://github.com/ECP-WarpX/WarpX/pull/1600), where the test was originally introduced.

-> @RemiLehe & @EZoni : ok to update checksum

[ax3l]

I did reset the three checksums with AMReX from a week ago. This passed locally.

After that, I pushed the latest AMReX change from this week on the PR.