While scanning other AMD 5950X setups, I noticed that computer 1024859 had RUN times and CPU times rather confusing.

It looks like Run Time is great than CPU Time and I wonder how that can be when the same operating system is used.

Is this perhaps the result of running in a VM or something?

Thanks!

8-)

I run one task per core instead of multicore. My run times are about 9% greater than my cpu times. I wasn't exactly clear on your question though.

OH!!! Okay. I run MT 8 cores per...

I suppose if there are other things going on, the cores are obviously shared with other processes
so run time could/would/is greater.

Pardon my old man moment...

Obviously I need more coffee...

8-)

OH!!! Okay. I run MT 8 cores per...

I suppose if there are other things going on, the cores are obviously shared with other processes
so run time could/would/is greater.

Pardon my old man moment...

Obviously I need more coffee...

8-)

One thing to remember is none of these tasks need done by yesterday so if you are running the project https://wuprop.boinc-af.org/ which tracks the hours you put in in each of the apps you run then finishing the task asap isn't always the best way. ie these are my MilkyWay hours:

Milkyway@home Milkyway@home N-Body Simulation with Orbit Fitting 155,058.32
Milkyway@Home MilkyWay@Home N-Body Simulation 103,409.68
Milkyway@Home MilkyWay@Home 100,411.85

And yes I really do have 100K hours in each of the MilkyWay apps they have had over the years I have been crunching here.

I see something even more weird -- I am running nbody_orbit_fitting on an Intel Windows 11 machine supposedly using two cpus per task, yet Task Master shows that they are using zero cpu time! But I am getting credit . . . I would attach a screen shot of the Task Master entry but I can't figure out how without first uploading it to a web page . . .

I see something even more weird -- I am running nbody_orbit_fitting on an Intel Windows 11 machine supposedly using two cpus per task, yet Task Master shows that they are using zero cpu time! But I am getting credit . . . I would attach a screen shot of the Task Master entry but I can't figure out how without first uploading it to a web page . . .

Yes you need to upload it to a site of your own and then link it here, alot of people use https://www.flickr.com/ but there are plenty of others too.

You said they show zero cpu time is that for the entire task or just at the moment you are looking at them?

It looks like on one of your pc's your tasks is taking around 11K to 13K seconds per task. On another pc they are taking in the 4K to 5K seconds range. While a 3rd pc is taking between 2.5K and 6k per task.

I am running nbody_orbit_fitting on an Intel Windows 11 machine supposedly using two cpus per task, yet Task Master shows that they are using zero cpu time!

This and in particular this computers seem to be highly overcommitted as the runtimes are close to CPU time, on 976179 the run time is usually even higher than CPU time. When running on two cores, the CPU time should be close to 2x run time. So yes, there are likely many periods of time, when the Milkyway does not get any CPU time. Check what else is using CPU and either fix it if it's some kind of issue with another process or limit the number of CPUs used by BOINC.

---

Zero CPU time showing up in Task Manager continually. The problem appears to be with my older Intel CPU -- single CPU with four "threads". I have two newer Intel machines, one with a quad CPU and the other with a hex CPU and they run milkyway tasks requiring two CPUs just fine. Last night I tried running tasks in the older machine requiring a single CPU -- they ran just fine and CPU usage showed up normally in Task Manager. Go figure . . .