Results in the computation of neutral axis depth (see test_fiber_section_prop.py) are different in different machines (difference about 10%). This is an error that we must catch but it will be necessary to have some way to transfer partial results between machines quickly and accurately. This way we can
compare results step by step to find the problem. A possible solution is to use 'pickle' with XC classes: https://www.boost.org/doc/libs/1_46_1/libs/python/doc/v2/pickle.html
lcpt
commented
2 years ago
Results in the computation of neutral axis depth (see test_fiber_section_prop.py) are different in different machines (difference about 10%). This is an error that we must catch but it will be necessary to have some way to transfer partial results between machines quickly and accurately. This way we can compare results step by step to find the problem. A possible solution is to use 'pickle' with XC classes: https://www.boost.org/doc/libs/1_46_1/libs/python/doc/v2/pickle.html