How Diagnostics work?

[KMalaravan]

For particle full particle diagnostics, e.g position, is the data outputted the average position of all the macroparticles loaded into the simulation? Is there any way of accessing the data of multiple particles from the input files, or is it only possible through accessing Particle wrappers in Python?

[RemiLehe]

Hi,

The full particle diagnostic output the position of each macroparticle (not the average position over all macroparticles), into an openPMD file (or a plotfile depending on the output file format that you selected). You can then open the file with Python postprocessing tools to extracting the position of individual macroparticles.

More info here:
https://warpx.readthedocs.io/en/latest/dataanalysis/formats.html
For openPMD files: https://warpx.readthedocs.io/en/latest/dataanalysis/openpmdviewer.html
For plotfile files: https://warpx.readthedocs.io/en/latest/dataanalysis/yt.html

[KMalaravan]

In terms of extending the simulation and getting data from each timestep, how are particle ids sorted. Is it in terms of momentum or positions of the particles, etc. I was plotting the position of 2 particles in a simulation over time, but there seems to be some id switching that I cannot explain otherwise.

[RemiLehe]

Thanks for checking.
In the above answer, when mention the id, do you mean:

• the index of a given macroparticle in the array of macroparticles (which indeed changes from timestep to timestep, since we periodically reshuffle macroparticles in the array to optimize memory access -- this might correspond to the "id switching" that you mentioned.)
• or do you mean the actual particle ID, which can be extract in openPMD-viewer for instance with ts.get_particle(['id'], ...) ; by extracting the id, you can track a given macroparticle consistently across timestep, even though its index in the array is changing. openPMD-viewer also has convenient method to do so, see:
  https://openpmd-viewer.readthedocs.io/en/latest/tutorials/4_Particle_selection.html