Two unusual behaviors in AWAKE simulation: CKC+PML tail filamentation and PSATD+periodic longitudinal modulation pattern

[556482]

Hi everyone,

I'm relatively new to WarpX and I'm trying to reproduce the AWAKE experimental results. I've encountered two different puzzling behaviors depending on the solver and boundary conditions, and I would appreciate any insights.

Setup:
Drive beam injected from an external file (half-Gaussian in z)
Moving window enabled

Case A: CKC solver + PML in all directions
The tail of the drive bunch develops violent oscillations and strong filamentation.

Case B: PSATD solver + PML in z-direction, periodic in x and y directions
The drive bunch does exhibit self-modulation, but the longitudinal microbunch pattern shows a "strong at both ends, weak in the middle" profile along the bunch.

My questions:
Has anyone observed similar tail filamentation with CKC and PML?
Has anyone seen this kind of longitudinal self-modulation pattern (strong head and tail, weak middle) when using PSATD with transverse periodic boundaries?
What might be causing these behaviors?
I'd be very grateful for any suggestions. Thank you!

CKC+PML PSATD+periodic

[RemiLehe]

Thanks for reporting these issues.
Would you be able to share the input scripts with which you observed this?