Waiting time in liphpc should be limited

[amjjbonvin]

Is your feature request related to a problem? Please describe.

In HPC mode (and also local mode with libparallel) the waiting time between checks of the completion of model is automatically adjusted during a run. For the HPC mode, in case of heavy cluster load this can lead to very long waiting times (I have seen 2000 seconds) which are not efficient as even if the jobs start running and complete fast, the scheduler will wait for a long time before the next check.

Describe the solution you'd like

Put a maximum to the automatically adjusted waiting time, e.g. 600 seconds.