LHC@home

I got one task that has an hour to go that shows in the main page in the status column Postponed: b.
There is nothing unusual in the event log.
Is this something like the Atlas problems with sensitivity or something?
Any ideas?

It could have been from running out of memory.
Suspend it and reboot the pc and then start the task again and check your Windows task manager *Performance/Memory* and see if it is close to using 100%

I see you ran 7 cores on a single task and it can be done if you have more ram than you have but less than 16GB can work with Theory tasks but not with others such as Atlas.

it is better to run 2-core tasks

It could have been from running out of memory.
Suspend it and reboot the pc and then start the task again and check your Windows task manager *Performance/Memory* and see if it is close to using 100%

I see you ran 7 cores on a single task and it can be done if you have more ram than you have but less than 16GB can work with Theory tasks but not with others such as Atlas.

it is better to run 2-core tasks

You and another guy over in Atlas talk about all these custom modifications.
I have no idea how to specific project limit cores and run times and make a second client to run Atlas on.
For Atlas was referred to a page on BOINC, but I haven't really dug into it due to time constraints.
For your idea of limiting cores, how do you do that?
Memory? Not a problem. I can run Atlas and do web stuff and still have 10% or more left over while running stuff like Asteroids, Universe and Milkyway all at the same time and web browsing.
But a reboot sounds like a goo idea to start.
----------
Reboot worked. Odd that this happens. Not all the time, but every now and then. Is it a windows thing or a BOINC thing?

For your idea of limiting cores, how do you do that?

The general idea is to create an "app_config.xml" file, using the "save as" function in a text editor to create the .xml extension.
Then you place the file in the LHC project folder, and read in the file or restart BOINC (or just reboot).

Here is an example, though I don't know if all the memory values are still correct:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4213&postid=29873#29873

For Native Theory, I would try the following. The numbers should be about right, though I have not done it myself yet.
This is set for two cores per work unit, with a maximum of four Theory work units and eight maximum for all the LHC projects.
Change them as you wish.

<app_config>

<app>
<name>TheoryN</name>
<max_concurrent>4</max_concurrent>
</app>

<app_version>
<app_name>TheoryN</app_name>
<plan_class>native_theory</plan_class>
<avg_ncpus>2</avg_ncpus>
<cmdline>--nthreads 2 --memory_size_mb 1500</cmdline>
</app_version>

<project_max_concurrent>8</project_max_concurrent> 

<report_results_immediately/>

</app_config>

If this doesn't work, computezrmle is the expert, and will be able to fix it up.

greg_be has a Windows machine, so can't run theory_native.

Reduce the limits in your preferences https://lhcathome.cern.ch/lhcathome/prefs.php?subset=project
and set Max # jobs to 7 and Max # CPUs to 1.

btw: VBox v 6.0.4 is much more critical to postpone tasks (missing heartbeat issue)
Try latest version 5: VBox version 5.2.26

greg_be has a Windows machine, so can't run theory_native.

I overlooked that. It will definitely limit his options.