Spectral solver with PML Boundaries and moving window about warpx HOT 3 OPEN

Thanks for reporting this issue, @sheroytata, and sorry for the late reply!

Regarding the reflections: This indeed seems strange ; could you share the input script for which you observed this issue, so that we could try to reproduce it locally?

Regarding restarting from a checkpoint with a different solver: similarly could you post the input script that you used for the restart?

Thanks!

from warpx.

Attached is a reduced version of the input script without particles and a smaller spot(so the laser increases in size faster) with the psatd solver. At iteration 17500 reflections seem to become visible in the laser field on the same color bar.

Attached is an example

and the input file
inputs_3d.txt

To restart, we just change the solver to ckc and change the boundaries to all PML
Say we restart from iteration 5000 and try to propagate the laser

Laser field after restarting from checkpoint + 100 steps

Also Attched is the input file
inputs_3d_checkpointRestart.txt

from warpx.

I recompiled warpx 23.05 and it seems fine with the spectral solver, the same reflections do not show up.
Thanks for the update to the spack build.

Although 23.06 has the same issue that I reported earlier, restart from checkpoint seems to loose particles at the back of the box. Maybe some MPI ranks just do not output data for particles.
#3203

Thanks

from warpx.