November 8, 2024 - Collaboration meeting

[juanfem]

Discussion topics:

• Cavity modeling. ESS doing benchmarking of fieldmap model against TraceWin for ESS SC cavities. It was found that TraceWin handling of 1D fieldmaps has less accurate transverse focusing than PyORBIT. Tracewin 3D fieldmaps agree better with PyORBIT 1D fieldmaps. No experimental data is available.
  Andrei offered to present SNS data from scanning superconducting cavities in our next meeting.
  Austin raised an issue with space-charge computations for long bunches in circular machines. To understand that problem, he will have a student doing space-charge studies with PyORBIT.
• En-Chuan brought up issues with Dual-Harmonic cavities. Andrei mentioned that SNS is not using that class and that it was developed by GSI.
• Bunch interface. It was discussed that currently users need to load particles one by one to a Bunch object, including dealing with MPI. Loading and saving bunch objects this way has bad performance. This led to 2 discussions: one to support a binary format for the bunch and another one to support loading a saving a bunch from a list or array. Andrei and Sasha commented that in the past they avoided adding numpy as a dependency due to the breaking changes to their ABI.
• Documentation about extension modules. We need to write documentation for 2 cases: how to add a new extension module to the PyORBIT core and how to create an external package including an extension module (e.g., laser stripping).

The next meeting will occur on December 13, 15:00 UTC+1, 9:00 UTC-4.