LHC@home

My machine is running the Atlas simulation wu.

It's currently taking around an hour to do one second's worth of processing and has been running for 2 days 14 hours (the estimated time was 6 hours 24.). It's in danger now of missing its deadline.

I have another 4 of these work units queued, all estimated at 6 hours 24 and I'm not at all confident any of them will meet their deadline.

Anyone else seen this?

The native ATLAS are running OK for me. But you seem to be on VirtualBox, and have only 7.5 GB of memory.
That might be enough to run one, with nothing else. But you are probably swapping out to disk most of the time.

Hmm....

This morning this particular Atlas simulation was sitting at "running" with 00:00:00 remaining and 99.999% complete.. three of the other Atlas simulations I had queued had run overnight and completed, so it seems just this one is having problems (it took 3 days 09:58:15 to get to that stage).

It's passed it's deadline but I'll think I'll leave it a day or two to see if it properly completes and uploads its results.

btw I'm running BOINC on a real Linux machine with "swappiness" set fairly low so it doesn't swap out to disk very often, so this isn't a virtual machine. The simulation is using vbox... I don't really understand how or why it does that.

Sam

All LHC subprojects (except SixTrack and special native tasks) use a very complex software environment based on ScientificLinux or CentOS.
They run in a VirtualBox VM to hide the complexity from the users and to avoid conflicts with the host OS.

Since host and guest are strictly separated they can't share resources like the available RAM.
Your computer (https://lhcathome.cern.ch/lhcathome/show_host_detail.php?hostid=10682666) reports only 7.5 GB RAM.
The recommended minimum for ATLAS and CMS is 16 GB.

SixTrack and Theory should run fine on this computer.
You may consider to set your prefs accordingly:
https://lhcathome.cern.ch/lhcathome/prefs.php?subset=project