numba.core.errors.LoweringError: Failed in nopython mode pipeline (step: native lowering)
cannot store {{i8*, i8*, i64, i64, double*, [1 x i64], [1 x i64]}, i1} to {{i8*, i8*, i64, i64, double*, [1 x i64], [1 x i64]}, i8}*: mismatching types
File "LMAHeureuxPorosityDiffV2.py", line 312:
def pde_rhs(state_data, t=0):
<source elided>
for i in prange(no_depths):
^
This error did not occur in my development folder, only in a new folder after a fresh git clone. I could not determine why, but the solution was to remove the parallellization of the jitted computation of the rhs. And it became slightly faster, from 2m09s to 1m46s, oddly enough. So instead of investigating the origin of this error, I fixed it by removing the parallellization. And got a speedup.
This branch was created to fix the Numba error
This error did not occur in my development folder, only in a new folder after a fresh git clone. I could not determine why, but the solution was to remove the parallellization of the jitted computation of the rhs. And it became slightly faster, from 2m09s to 1m46s, oddly enough. So instead of investigating the origin of this error, I fixed it by removing the parallellization. And got a speedup.