Pause/Restore simulation after network failure (in Parallel)

[Anirban Basak]

Hi,

           I have two 12 core machine in our lab and a 80 core SUN server in other lab. The license server is one of the two machine in our lab. I am trying to run a simulation over the SUN machine. But we have some power failures and the simulation gets killed without the license. Is there any way to pause the simulation automatically (when network fails/no license) and resume it again automatically (when power comes back)? The simulation will be waiting for the license to become available again and not get killed (like Synopsys TCAD Sentaurus).

Thanks

[Anders Blom]

There is the checkpoint file, that's the only option currently.

The license has a 5-minute time-out, beyond that I can't say we have any options available.

[Anirban Basak]

Hmm. Can we not increase the license time-out to say 1 hr, so that the simulation waits for sufficiently long time before it stops?

[Anders Blom]

I'm afraid not, and besides it would not perhaps be the appropriate solution. The license time-out doesn't pause the calculation, it just continues in the absence of a license connection, until it can be established again.

We can offer you a possibility to host the license closer to where the calculations run - after all, the machine which runs the job has power, so that would be a more suitable location for the license server. Please contact your sales rep for discussions about how to change the license server.