Hello Yosef,
I should have clarified this to begin with, I was using a single processor.
Well in that case my comment clearly does not apply.
It seems that somehow the extent of the grid array is lost. The grid, as seen by AEILocalInterp is only 9 elements long.
This may be unrelated: but are you using a development version of the flesh? I fixed a bug in the parameter file parser that affects things like:
size=\ (NullGrid::N_ang_pts_inside_eq+2*(NullGrid::N_ang_ev_outside_eq+NullGrid::N_ang_stencil_size)),\ (NullGrid::N_ang_pts_inside_eq+2*(NullGrid::N_ang_ev_outside_eq+NullGrid::N_ang_stencil_size))
in interface.ccl (the fix is in 635f90e "Cactus: fix access to deallocated temporary variable in piraha"). This would affect the size of the array in memory as well though, so I would expect it to fail before.
Otherwise, I am as baffled as you are.
Yours, Roland
users@lists.einsteintoolkit.org