Say, I have a number of units to be evaluated, units evaluated one by one. Under the term ‘evaluation’ I understand the following:
- Run the kernel for a single unit.
- Wait until kernel is finished.
- Check the result.
The problem is in step (2) as I want not only to wait but to terminate the kernel if it works for too long (in two words: if unit can’t be evaluated in less then, say, 0.5 seconds - then it is considered as ‘bad’ and I don’t want to waste time on it’s full evaluation).
Is to possible to implement such timing ?
Thanks in advance!