Open whl64 opened 1 year ago
@whl64 This solver is currently exercised well using a Boris push, direct current deposition, and particle shape factor of 1. If you look in Examples/Tests/magnetostatic_eb there are some example inputs with an existing beam inside of a beampipe.
algo.particle_pusher = boris algo.field_gathering = momentum-conserving algo.particle_shape = 1 algo.current_deposition = direct
I adjusted these parameters in your input script and it runs for me.
Hopefully this can get you started with the solver. It currently only supports homogeneous embedded boundary conditions, so it can't easily support something like an embedded conductor cathode. It works well for beams moving transverse to the conducting surfaces.
Thank you for reporting this.
Thanks, that's good to know. Is it also possible to use the relativistic solver with embedded boundaries?
@whl64 I have DM'd you in gitter. I would like to keep this issue clean so when I commit changes and refer to this issue it isn't muddled by side conversation.
The error message is "Error in `warpx.3d.MPI.OMP.DP.PDP.OPMD.EB.QED': corrupted size vs. prev_size: 0x000000014b90fc50". When using a Gaussian injection, the labframe solver, or a binary compiled without EB support, the error does not occur. I have attached an example input file that causes the error. eb_emstatic_test.txt