LHC@home

I have some running at -dev but that site needs to be poked with a stick since it is just blank pages for everything so we can't see what is going on.

Still waiting for lhc-dev website....

All the tasks I've gotten on Linux PCs have been single threaded even with it set to be 4 threads in LHC preferences. The lone task I've completed in Win10 was mt. Both say Setting CPU Count for VM. (4)

Dev site seems to be missing but services must be up as I can download a task.

All the tasks I've gotten on Linux PCs have been single threaded ...

This can't be verified since your computers are hidden.
And, no, please don't post the logs.

CERN Grafana shows that since yesterday all CMS jobs were 4-core jobs.

https://lhcathome.cern.ch/lhcathome/result.php?resultid=418694926

Run time 46 min 22 sec
CPU time 33 min 47 sec

Ya don't want logs? Posting task links is not useful for history as task data is purged. I've often come to BOINC forums and been unable to verify if I am having the same issue as described as only a link was posted. It also doesn't help when searching either. For this PC, I was able to fix the vbox issue on this PC yesterday by searching LHC forums with the error. This was only possible because someone posted their log.

Your VMs are configured to use 4 cores.

Check the consoles of a VM that is active for at least 30-35 mins (setup phase must be finished).
console 3: should show 4 instances of "cmsExternalGene" each using close to 100% CPU
console 4: should show "WMAgent Run the job ..." as last line
console 5: should show a couple of lines like "INFO:root:PSS: ... PCPU: 15.0; PMEM 4.4"

If the VM finishes without all of that the job did not fully start.
The low CPU time compared to the walltime suggests the same.
This happens sometimes if IP connections get lost.
Check if your local firewall blocks any connections initiated by the VMs.

Also check the VM internal logs.
Stderr.txt reports the port where they can be found, like:
Detected: Web Application Enabled (http://localhost:42069)
Type this in a browser on the host running the VM:
http://localhost:42069/logs

bonjour,
j'ai reçu 112 taches cms avec un temps d'execution de 17h50 avec 4 threads.
est ce normal?
hello,
I received 112 cms tasks with a time of 17h50 with 4 threads.
Is this normal?

---

Not really but it happens when the estimated compute performance of your computer is high vs the estimated compute for the WU.

On your side you can run the benchmarks in BOINC, for the estimated compute this is set by the project, you can overide it by editing the WU, for me this is a bit too much effort but others have done it.

CMS tasks running scientific jobs usually finish between 12-18 h.
The BOINC benchmarks are not relevant since each VM runs internal benchmarks that are not reported back to BOINC.
Based on those "glidein" tries to calculate if there's enough time left to request another job.

Each pause/resume initiated by BOINC disturbs this calculation and may result in loosing the last scientific job (18 h are a hard limit).
Since this is also not reported back to BOINC it is recommended to avoid pause/resume cycles.

The short runtimes during the last few weeks were tasks without any scientific job.
Just empty envelope VMs doing a basic setup and then shut down after a timeout.

To be sure a scientific job is running, check the consoles and logs as described here:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=6265&postid=51364

I noticed for the first time, that 1 CMS-job is now processing 120,000 events in stead of the 100,000 in previous batches.
Is this the new default number of events?

I noticed for the first time, that 1 CMS-job is now processing 120,000 events in stead of the 100,000 in previous batches.
Is this the new default number of events?

Yes, I changed that a while back to adjust job timing.

---

The project is now delivering a new version of CMS application (70.60).

---

Same as before, except that this app_version uses vboxwrapper 26208 like Theory published earlier this afternoon.