Axel Huebl
Axel Huebl
Another benefit of splitting off the compute part would be that we can compile each element in their own `.cpp` files. Currently, they all end up in the same TU,...
Another aspect we could keep in mind that we can implement in the backend: we add more and more fallbacks to drifts for elements with zero fields. Currently, we add...
@titoiride can you potentially test this PR with your data, too? :) Please restart from a checkpoint that was written from a simulation that ran from the beginning (not a...
@RemiLehe > In principle, warpx.getmoving_window_x() should be able to work with starting/stopping moving window. To avoid duplicating code (as you pointed out), should we remove the function warpx.getmoving_window_x() (since it...
## First backtrace (with load balancing) One segfault points here (RZ Spectral solver): 3a. https://github.com/AMReX-Codes/amrex/blob/25.11/Src/Base/AMReX_BaseFab.H#L377 3b. https://github.com/BLAST-WarpX/warpx/blob/25.11/Source/FieldSolver/SpectralSolver/SpectralFieldDataRZ.cpp#L247 The `DampPML` segfault points here: 3a. https://github.com/AMReX-Codes/amrex/blob/25.11/Src/Base/AMReX_FabArray.H#L569 3b. https://github.com/BLAST-WarpX/warpx/blob/25.11/Source/BoundaryConditions/PML_RZ.cpp#L94 ## Second backtrace (w/o...
## RZ Spectral Solver & Load Balancing For the first backtrace, I wonder if we get with load balancing out of sync between fields in the RZ spectral solver. Are...
# CurrentDeposition (no LB) We lowered the CFL from `0.999` to `0.95` and it seems to get further for the deposition.
@titoiride dang. As a crude work-around, try lowering it further maybe, to 0.9 or even lower? Does adding the `AMREX_ALWAYS_ASSERT_WITH_MESSAGE` point to particles that are out of bounds?
Just to [close the loop on this](https://github.com/BLAST-WarpX/warpx/issues/6373#issuecomment-3514223808): @dpgrote commented on Slack: > @ax3l asked about the temporary arrays in the SpetralFieldDataRZ class. These will be regenerated during a load balance...
> https://github.com/BLAST-ImpactX/impactx/pull/964#issuecomment-3085567714 Fixed GPU compile :) This was just lacking a host device annotation in a function declaration.