Advanced search

Forums : Technical Support : camb_boinc2docker tasks aborted/computation error
Message board moderation

To post messages, you must log in.

AuthorMessage
jave200372

Send message
Joined: 28 Jan 08
Posts: 1
Credit: 303,007
RAC: 11
Message 23022 - Posted: 28 Apr 2022, 0:33:33 UTC

Hi
All my boinc2docker tasks are finishing with a computation error...

Thu 28 Apr 2022 08:38:55 | Cosmology@Home | Aborting task camb_boinc2docker_1206098_1650974314.514798_0: exceeded elapsed time limit 828.08 (86400.00G/104.34G)
Thu 28 Apr 2022 08:39:03 | Cosmology@Home | Computation for task camb_boinc2docker_1206098_1650974314.514798_0 finished
Thu 28 Apr 2022 08:39:03 | Cosmology@Home | Output file camb_boinc2docker_1206098_1650974314.514798_0_r2112742825_0.tgz for task camb_boinc2docker_1206098_1650974314.514798_0 absent


They typically abort after 11-13 minutes. Can anyone tell me what is going on?
ID: 23022 · Report as offensive     Reply Quote
Jim1348

Send message
Joined: 17 Nov 14
Posts: 134
Credit: 5,412,499
RAC: 17
Message 23023 - Posted: 28 Apr 2022, 0:42:20 UTC - in response to Message 23022.  

You are using BOINC 7.18.1. That has problems on VirtualBox projects, and some others.
https://boinc.berkeley.edu/forum_thread.php?id=14650

You probably got it from ppa:costamagnagianfranco/boinc, but that doesn't matter.
Try removing it (and remove the ppa).

Then, install the version that comes with Mint, which will probably be 7.16.6, which is what I get with Ubuntu. It should work.
ID: 23023 · Report as offensive     Reply Quote
Jonathan

Send message
Joined: 27 Sep 17
Posts: 185
Credit: 8,306,392
RAC: 5,376
Message 23027 - Posted: 29 Apr 2022, 4:01:23 UTC - in response to Message 23023.  

I checked a single work unit and you are assigning all 24 cores to a task.
Edit your cosmology@home preferences for Max # jobs & Max # CPUs

Max # CPUs should be 8 or less. I suggest 2 to 4. It controls the number of cpus assigned to a camb_boinc2docker task.

The tasks complete quite quickly with more cores than that and you spend a lot more time setting up and deleting the VM than doing work.

You will have to find the balance for your computer on how many concurrent work units at once and how many cores per work unit
ID: 23027 · Report as offensive     Reply Quote
[AF>Le_Pommier] Jerome_C2005

Send message
Joined: 4 Mar 13
Posts: 2
Credit: 1,005,192
RAC: 0
Message 23035 - Posted: 5 May 2022, 18:35:37 UTC
Last modified: 5 May 2022, 18:36:07 UTC

The tasks I got in my mac also failed, some failed quite quickly "on their own", and others I found them without any CPU usage after more than one hour running, I had to cancel them.

AFAIK I have latest version of VB and boinc for mac.

Maybe the problem has no link with the previous ?
ID: 23035 · Report as offensive     Reply Quote
Jonathan

Send message
Joined: 27 Sep 17
Posts: 185
Credit: 8,306,392
RAC: 5,376
Message 23036 - Posted: 6 May 2022, 1:59:34 UTC - in response to Message 23035.  

Do you have any other projects using VirtualBox and are they working? How many concurrent tasks are you running for Cosmology?
You can try setting your project preferences to do one task at a time. Max # of Jobs.
I took a look at a few of your tasks logs and I don't see anything obvious.
ID: 23036 · Report as offensive     Reply Quote
[AF>Le_Pommier] Jerome_C2005

Send message
Joined: 4 Mar 13
Posts: 2
Credit: 1,005,192
RAC: 0
Message 23037 - Posted: 6 May 2022, 6:11:04 UTC

It's been a (long) while i didn't do any cosmo but AFAIR it used to work with VB, i don't have any remaining setup to limit the task number (I have some with other projects)... But saying i never had any trouble with VB projets is untrue, i always have troubles...

I'll try to limit and see what happens.

Thanks
ID: 23037 · Report as offensive     Reply Quote

Forums : Technical Support : camb_boinc2docker tasks aborted/computation error