Joined: 6 Mar 15
My apologies if I've missed discussion of this in an existing thread...
I've noticed that there are often significant differences in a work unit's run time and CPU time. In longer running tasks, I've seen differences of more than 10K seconds. It seems like these come in waves and may or may not be related to particular batches of work.
For example, Task 27936510: The run times for my AMD-Linux host and my wingman's Wintel machine were about the same. However, my CPU time was ~11K seconds less.
I'm wondering if it might be Linux-related as I've only noticed this on my two Linux hosts. However, they run a lot more tasks than the one Wintel host I have running right now.
In what may be a separate issue, I've noticed that these tasks can go on for hours after the last checkpoint. But, they continue to progress and eventually finish.
Joined: 28 Feb 15
I'm wondering if it might be Linux-related as I've only noticed this on my two Linux hosts.
I have not noticed that on my Ubuntu machine. The CPU time always tracks the run time fairly closely.
However, Thyme Lawn posted an interesting one, but for a different reason (see his wingman). https://universeathome.pl/universe/workunit.php?wuid=12316088
And note that it was marked a "success", even though the stderr output indicated a problem. https://universeathome.pl/universe/result.php?resultid=27920271
So there are strange things going on, but I don't know what.