[ShawnL00]

Add options for source box parallelization for P2M and P2L expansions.

P2L:

• Each target box is a workgroup, and each work item handles the assigned source particles.
• Source boxes for a given target box are processed in sequential order.
• After processing all source points for a given target box, each work item sums the partial results for the assigned coefficients.

P2M

• Similar to P2L, each source box is a workgroup.