LHC@home

I was running Atlas from the original website and have recently changed over to LHC.

I have noticed that all the WU's on my macbook pro are failing. Not sure why, given that they were running fine when taken directly from Atlas.

Here is an example of one the WU's.
https://lhcathome.cern.ch/lhcathome/result.php?resultid=129931920

Take this app_config.xml in the Boinc folder of the project:

lhcathome.cern.ch_lhcathome

<app_config>
<app>
<name>ATLAS</name>
<max_concurrent>1</max_concurrent>
</app>
<app_version>
<app_name>ATLAS</app_name>
<avg_ncpus>3</avg_ncpus>
<plan_class>vbox64_mt_mcore_atlas</plan_class>
<cmdline>--memory_size_mb 5000</cmdline>
</app_version>
</app_config>

avg_ncpus is the max CPUs you want to take part.

Memory_size_mb must be greater or equal 4.400.

At first glance it looks like a segmentation fault in ATLAS code, i.e. a faulty WU, but all your WUs fail like this so it's not WU-specific.

I looked deeper into the logs and there is a problem reading database information. Maybe you have a firewall blocking the connection? Although the connections required in LHC are exactly the same as in the old ATLAS@Home.

When I changed my setting from 1 core WUs (running with 3400 MB) to 2 core WUs I noticed similar error messages in stderr.txt and the WUs didn´t validate.
Even 4480 MB were not enough.
With 5120 MB the most recent WU is now running.

I guess it´s a RAM issue instead of a firewall issue.

Guys,

Thank you for the info/assistance. I would advise that I updated my VM & BOINC on my machine and it now appears to be running OK. However, it has picked up one of the long WU's and is currently at day 4.

This Apple produce a lot of unresolved ATLAS-Tasks:

https://lhcathome.cern.ch/lhcathome/results.php?hostid=10467467

So, other Computer become those Tasks to finish them.