| 7 | | However, the case reported here is different. In this case, there is no division by zero _in the torn equations_. What happens instead is that _the residual equations become singular_ because of a bad choice of tearing variables. I guess this may be more difficult to detect. What do you think? |
| | 7 | However, the case reported here is different. In this case, there is no division by zero _in the torn equations_. What happens instead is that _the residual equations become singular_ because of a bad choice of tearing variables. I guess this may be more difficult to detect. |
| | 8 | |
| | 9 | Since the singularity is structural, it could be detected at compile-time and invalidate the proposed tearing that lead to it, triggering a second (or a third) try. |
| | 10 | |
| | 11 | What do you think? |
