The previous implementation was quite slow in filling up the remaining_femelements list. The new implementation fixes this by using numpy for the heavy work.
The previous implementation was quite slow in filling up the remaining_femelements list. The new implementation fixes this by using numpy for the heavy work.