2D ellipsoid space charge calculator

[austin-hoover]

Suggestion to create a 2D version of UniformEllipsoidFieldCalculator for simulations with long bunches in rings.

[shishlo]

I agree. That is a great idea. The set of 2D Ellipsoids with/or without boundaries will give us a fast tool for ring dynamics calculations. Parallelization scheme should be more flexible and faster then FFT based algorithms.

[austin-hoover]

Implemented in PR #143, mostly by AI. Looks like it's giving correct results, comparing to 2D PIC with KV distribution (1 ellipse) or Gaussian distribution (5-10 ellipses) for long coasting bunches.

• No boundaries conditions were implemented.
• Have not tested with MPI.
• Do we want to weight transverse kicks by longitudinal density, like in SC2p5D?
• Similarly, do we want to add a node under SCSliceBySlice2D? This would run a separate ellipse solver in each longitudinal slice.
• I renamed classes to follow same convention as other space charge nodes.