Dear all,
following up on last week's questions regarding a tracer particle implementation into an evolution code I am having problems with the function *CCTK_**InterpGridArrays*. In general, my implementation follows the GRHydro_particle approach (in C++), and is working in some cases. However, at some interpolation points, the interpolation routine returns nans/infs. This error occurs mainly for the interpolation points in refined regions and changes when I change buffer zones, grid refinement or scheduling options.
Currently, I am not finding the explicit problem, and, I observe that my tracer particles are evolved fine in the less refined regions (e.g., further away from the origin). For example, I may set up 500 particles and in the first evolution step the interpolation produces 112 nans/infs. Throughout the further evolution, the remaining 388 particles (the ones the most far away from the origin) are evolved fine. Therefore, I assume I am missing something in my understanding of the interpolation routines. Has somebody encountered similar problems or an idea where I could direct my debugging to overcome the interpolations to NaNs/Infs?
Thank you very much and my best from Valencia
Jens