LHC@home

Just added ATLAS Simulation. Downloaded the latest version of VM box 6.1 and downloaded a few Atlas jobs.
I suspended all but 1 job and its started and "seems" to be running?. It is doing about 6.5%/hr.

I notice that the cpu load is averaging about 60%. Is this normal?. All other other jobs from other projects seem to run 100%. Even the six track runs max speed.

The normal windoz task manager show very little cpu load.

Is this all normal?
Ta
Nairb

Yes it is normal
They run as long as 1 day 8 hours for a single core Atlas

Thanks for the info. Its running all 4 cpu's and still at about 60%.

Its been running nearly 24hrs and its at 79%. Somewhat longer than the 15 hrs stated in boinc.

It looks like running another 12 hrs at least.

Still all normal??
Ta
Nairb

If you are looking the progress in Boinc Manager, then forget it. VirtualBox tasks do not communicate their progress back to Boinc Manager. Manager is showing just a simulated progress that is not real. You can see the actual progress only within the VirtualBox task. See more here: https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4161&postid=29359#29359

---

So, in the VirtualBox manager I press the green "show" icon. I get a black window which I can try and log on. If I select "machine from the menu and then session information from the drop down menu I get another screen showing cpu load at about 50% average. But I dont know if that is all four cpu's or just 1.
From Boinc if I select properties from the running job is says 7.30mins of cpu time and 4 seconds CPU time since checkpoint. With an Elapsed time of 18.36 hrs.

I think I have opened all the correct port numbers.

I dont think this is working. And I dont think it ever has on this win10 machine for some reason.

I ran the V-CPU from LeoMoon and have 2 green boxes.

I bet if I leave this running it would take 3+ days if it dosen't fail first

Your start is correct. In the black window press Alt+F2 to get a window of how much work has been done so far and how long it is estimated to still take. With Alt+F3 you get linux top window that shows the CPU and memory usage. The main process where calculations are done is athena.py (there should be one of these per used cpu core). You'll get different terminal windows with Alt+function key, the most useful ones are the Alt+F2 and Alt+F3.

---

Right I have a major step forward. For some reason the alt f3/f2 gave the login prompt (again) but this time I used atlas login (again) and for the first time ever I was logged in. Hurray. Now I am more familiar with linux so first i found out that it does see 4 cpus.
Top produced some interesting info. A quick grep with athena and nothing was found. All cpu's are idle. The id in top is 99.9%. Nearly all memory is free.
So it would seem that nothing is being processed. There are 4 cpu's in the top list.

Is there a log somewhere that might show why its failed or not processing? or failed to start.

Ta
Nairb

The trouble starts with the fact that your computer reports less than 8 GB RAM in connection with the fact that VirtualBox locks 6.6 GB for the VM:
https://lhcathome.cern.ch/lhcathome/show_host_detail.php?hostid=10495038
https://lhcathome.cern.ch/lhcathome/result.php?resultid=352727533

2022-04-22 17:05:13 (4632): Setting Memory Size for VM. (6600MB)

This is the reason why your BOINC client switches between "paused" and "running".
At the end the VM crashes and leaves too much garbage which needs to be cleaned before the next try:

2022-04-23 03:57:32 (21840): 
   NOTE: VM session lock error encountered.
 		    BOINC will be notified that it needs to clean up the environment.
 		    This might be a temporary problem and so this job will be rescheduled for another time.

To solve this:
1. Set BOINC to "don't request new tasks"
2. Cancel all ATLAS tasks (as they are all 4-core tasks)
3. Report them back to the project
4. Shut down BOINC
5. Open the VirtualBox Manager and remove all broken ATLAS VMs (if any are shown)
6. From the menu call the media manager and remove all broken "vm_image.vdi" files (if any)
7. Limit your LHC@home preferences to max 2 cores: https://lhcathome.cern.ch/lhcathome/prefs.php?subset=project
8. Restart BOINC
9. Resume work fetch

Yes that was the real problem - not enough memory. I followed the checklist but still needed to stop a couple of other processes to get enough memory for the vbox to work. Even the F2/F3 keys work now. I get 2 athema.py processes running and it has completed 10 of 200 events so far.

Its plain I will have to increase the amount of memory on this machine if I want to run atlas on a regular basis.

Thanks to all who have helped. I would have probably given up otherwise. I will try getting the I7 machine set-up with vbox but had better wait till the cpdn jobs have finished first.

Thanks
Nairb

I will have to increase the amount of memory on this machine...

Even then it's not recommended to run VMs with more than 2 cores each (on this machine).
See: https://forums.virtualbox.org/viewtopic.php?f=35&t=77413

Hurrrray its done an Atlas w/u and its been Completed and validated. The smart move now is to let it complete the w/u left and set up the desktop with vbox and run Atlas from there since its always connected to the internet. This laptop is better with the six track w/u.

But at least its worked for once.

Ta
Nairb