Calculates exponential propagators. Syntax:
returns exp(-i*L*t). The following calculation methods are supported:
'cpu' - Taylor series with scaling and squaring on CPU, spmd parallel if possible
'gpu' - Taylor series with scaling and squaring on GPU
The propagator calculation method is chosen by setting the sys.enable parameter during the call to create.m function.
Notes: we did have Chebyshev and Newton series here at one point, as well as the Pade method. None of them has lived up to their marketing.