LHC@home

I was looking at CPU time in Boinc Stats and in the Slot that ATLAS is using.
I see only a few seconds of CPU time being used every so many minutes/hours.

So my question is this: If it is not reporting CPU usage time, then what is it doing with the CPU time it is using and why does it not report what it is doing as CPU usage time?

Where do I see the check points for work done after 8 seconds of CPU time is used? I looked at the STDERR text and I don't really see any other information after 8 seconds.

Can someone explain? I am just curious why ATLAS reports CPU time differently than all the other BOINC tasks I run.

I was looking at CPU time in Boinc Stats and in the Slot that ATLAS is using.
I see only a few seconds of CPU time being used every so many minutes/hours.

Look at your results. They are all errors. It is possible that you need more memory. ATLAS take more than Theory.

I am having problems on native ATLAS though. It could be that they need to fix something.

I was looking at CPU time in Boinc Stats and in the Slot that ATLAS is using.
I see only a few seconds of CPU time being used every so many minutes/hours.

Look at your results. They are all errors. It is possible that you need more memory. ATLAS take more than Theory.

I am having problems on native ATLAS though. It could be that they need to fix something.

What I am seeing is VM unable to work due to something else locking up the slot that BOINC decides should be used. Which is odd, because I have seen VM tasks all working together ok.
There was a time that VBOX was having issues and needed to be cleaned up.

I am not sure what could be causing the memory problem though....I run FAH in addition to stuff like Prime Grid and Rosetta and others. I had a look right now at whats going on..I am 48% of memory running FAH (maybe I should stop running this now that I have tuned up the way BOINC uses my system for my lower end GPU card)

But that is not what I came here to discuss. I came here to discuss why even though the task does complete normally, it shows just 8 seconds of CPU time out of 1.5 days run time (being stopped for other tasks after 8 hrs and them taking 8 hrs). In all my crunching and having BOINC Tasks, I have never see ATLAS use more than 8 seconds of CPU time no matter if it errors or completes ok. So again, the question, what does ATLAS due with the CPU's outside of the 8 seconds reported CPU time?

But that is not what I came here to discuss. I came here to discuss why even though the task does complete normally, it shows just 8 seconds of CPU time out of 1.5 days run time (being stopped for other tasks after 8 hrs and them taking 8 hrs). In all my crunching and having BOINC Tasks, I have never see ATLAS use more than 8 seconds of CPU time no matter if it errors or completes ok. So again, the question, what does ATLAS due with the CPU's outside of the 8 seconds reported CPU time?

That is what caught my eye, though it is probably for a different reason in native ATLAS. I get practically zero CPU usage too. I think it just means that the task does not start up.
Whether they are related is another matter.

But that is not what I came here to discuss. I came here to discuss why even though the task does complete normally, it shows just 8 seconds of CPU time out of 1.5 days run time (being stopped for other tasks after 8 hrs and them taking 8 hrs). In all my crunching and having BOINC Tasks, I have never see ATLAS use more than 8 seconds of CPU time no matter if it errors or completes ok. So again, the question, what does ATLAS due with the CPU's outside of the 8 seconds reported CPU time?

That is what caught my eye, though it is probably for a different reason in native ATLAS. I get practically zero CPU usage too. I think it just means that the task does not start up.
Whether they are related is another matter.

I have another thread open from earlier where a user talks about limiting ATLAS to 4 cores and 3 threads or something like that. Need to back and read that and maybe put that script into play.

https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=5226

BTW....here is what I mean about only a 8 seconds...
2019-12-24 13:15:11 (2752): Status Report: Elapsed Time: '133301.871103'
2019-12-24 13:15:11 (2752): Status Report: CPU Time: '8.328125'
2019-12-24 15:14:27 (2752): Status Report: Elapsed Time: '139302.555752'
2019-12-24 15:14:27 (2752): Status Report: CPU Time: '8.375000'
2019-12-24 17:07:45 (2752): Status Report: Elapsed Time: '145302.555752'
2019-12-24 17:07:45 (2752): Status Report: CPU Time: '8.390625'
2019-12-24 18:54:15 (2752): Status Report: Elapsed Time: '151302.555752'
2019-12-24 18:54:15 (2752): Status Report: CPU Time: '8.453125'
2019-12-24 20:34:21 (2752): Status Report: Elapsed Time: '157302.555752'
2019-12-24 20:34:21 (2752): Status Report: CPU Time: '8.484375'

How does it use just a "blip" of CPU time when it is running?

Using the 4 core script someone gave me. Now ATLAS is acting normal.
But it is still in the early stages of processing.
See how it looks in a few hours.