Problem with a task

[ChristianB]

Hi Roland,

this is a known issue with XXL tasks. Those tasks have no checkpointing capability and need a host that is not powered down during the whole computation period. How did this task affect your local network? You could use the computer as usual but shouldn't reboot as the computation would start at the beginning.

If the task restartet recently you should abort it as this workunit get's converted to the more stable VM application that enables checkpointing through Visualization technology. If it is still running at 100% I will increase the deadline and you can let it run further. There is already one successful result for this WU that took 514 hours on an Intel Xeon 2,4 GHz CPU. So I would estimate that you are roughly at 50% of total computation time. You can still abort the task if you think you need to reboot your computer within those timeframe.

As for your suggestions, we already took care of the missing checkpointing mechanism by using a VirtualBox VirtualMachine with our cmsearch VM application. Because of the undeterministic nature of the cmsearch application in general we can't estimate the time needed (and thus the client capacities) for those very long tasks accurately or increase the time needed for tasks (which would be great!).

BOINC 7.2.33 by itself is not offering any checkpointing capability. This is achieved by VirtualBox and our VirtualMachine application (which you don't have on this computer).

The task hasn't finished, it's just that it takes more time than it initially thought it would need.