grid_base warning about opendf HOT 3 CLOSED

The warning is caused by the fact that when there's more than one bosonic frequency, the shift of the fermionic GF is required, i.e. G(w+Omega, k+q). Since in this case w+Omega maybe outside of the grid this shift is not possible for some points at the end of the fermionic grid, instead the zero is returned. The warning is thrown to notify of this. So this is an intended behavior. It is not done in the release mode, because it's a costly range check.

from opendf.

Ok, makes sense. This is also most likely the source of the high frequency problem that we have in James' chi. Just to make sure again: even in release mode, zero is returned, not some undefined behavior?

from opendf.

Yes. It is zero. This could be avoided by doing FFT instead of a shift. But this requires attaching tails, which in generally we don't know for the vertex. Alright, closing this one.

from opendf.