Published comparisons of BEM with other methods.
We are aware of 2 published papers in which the BEM is compared with the FDM and FEM methods for electrostatic charge particle optics:
41. Comparison of FDM, FEM and BEM for electrostatic charged particle optics, by D Cubric, B Lencova, F H Read and J Zlamal, Nucl. Instr. Meth. A427, 357-362 (1999).
36. Comparison of Finite Difference, Finite Element and Boundary Element Methods for electrostatic charges particle optics, by D Cubric, B Lencova and F H Read, Electron Microscopy and Analysis 1997, Institute of Physics Conference Series 153, 91-94 (1997)
-These deal with non-space-charge simulations and are taken from the list in publications, which also contains many papers that demonstrate the high accuracy of the CPO programs in various types of simulation. In general the above two papers show that the BEM is usually about two orders of magnitude more accurate than the other methods, for the same computing time.
A paper that deals with space-charge simulations is:
Comparison of simulation codes for the beam dynamics of low-energy ions, by K. Hanke, S. Heising, G. Probert and R. Scrivens, Reviews of Scientific Instruments, vol 73, 783-785 (2002). See also http://cdsweb.cern.ch/record/527583/files/ps-2001-061.pdf. These authors find good agreement between CPO, KOBRA and IGUN, all of which simulate a real beam line in agreement with the experimental data (to within approximately 2% or less). No information is given on the computing times.
CPO Ltd would be grateful to be told of any other published comparisons.