Message boards :
ATLAS application :
no ATLAS tasks available
Message board moderation
Author | Message |
---|---|
Send message Joined: 18 Dec 15 Posts: 1686 Credit: 100,488,944 RAC: 104,589 |
I now changed with one of my PCs on which I have been crunching ATLAS for more than a year, from the ATLAS Project to the LHC@home Project. Although according to the Project Status page more than 200 ATLAS Tasks are shown as "unsent", BOINC tells me that none are available. Am I doing something wrong, or is the error somewhere else? |
Send message Joined: 24 Jul 16 Posts: 88 Credit: 239,917 RAC: 0 |
What are the parameters you put in your preferences account for : #max cpu and #max job? |
Send message Joined: 15 Nov 14 Posts: 602 Credit: 24,371,321 RAC: 0 |
Also, ATLAS is still in beta, so you must check "Run test applications?" in the preferences. |
Send message Joined: 18 Dec 15 Posts: 1686 Credit: 100,488,944 RAC: 104,589 |
What are the parameters you put in your preferences account for : 3 each. so far, on the ATLAS Project, I ran 3 tasks with 3 cores each (using a total of 9 cores). Is my assumption right that with Parameters 3 and 3 I will have same situation? |
Send message Joined: 18 Dec 15 Posts: 1686 Credit: 100,488,944 RAC: 104,589 |
Also, ATLAS is still in beta, so you must check "Run test applications?" in the preferences. if I read correctly in another thread of the message board, the test application are all "longrunners" with 10-fold crunching time - correct? |
Send message Joined: 18 Dec 16 Posts: 123 Credit: 37,495,365 RAC: 0 |
the test application are all "longrunners" with 10-fold crunching time - correct? Not all, it is a combination of "normal" and "long" runners. The normal ones have a run time similar to the tasks in ALTAS project. For the time being there is no mean to select normal only or long only tasks. We are the product of random evolution. |
Send message Joined: 18 Dec 15 Posts: 1686 Credit: 100,488,944 RAC: 104,589 |
it all works well now. New tasks were downloaded (so far, no "longrunners"). |
Send message Joined: 18 Dec 15 Posts: 1686 Credit: 100,488,944 RAC: 104,589 |
what I noticed now: in contrast to how it worked with the former ATLAS Project, now with the LHC@home Project no jobs beyond the preset number of jobs can be downloaded. In other words: before, for example I had 3 tasks running, and - depending of the computing preferences in BOINC - several additional jobs were "ready to start". Now, with LHC@home this is impossible. When I have 3 tasks running and push the "update" button in BOINC, it says "the computer has reached a limit on tasks in progress". Why so? |
Send message Joined: 18 Dec 16 Posts: 123 Credit: 37,495,365 RAC: 0 |
Erich, if you want to have, say, 6 tasks downloaded and only 3 running concurrently, you need: 1. To set Max # jobs to 6 in LHC@Home preferences. 2. To set max_concurrent to 3 in app_config.xml <app> <name>ATLAS</name> <max_concurrent>3</max_concurrent> </app> This should achieve what you are looking for. We are the product of random evolution. |
Send message Joined: 18 Dec 15 Posts: 1686 Credit: 100,488,944 RAC: 104,589 |
This should achieve what you are looking for. Thanks, Herve - it does :-) |
Send message Joined: 13 May 14 Posts: 387 Credit: 15,314,184 RAC: 0 |
We should have more tasks later today (the normal ones, not the longrunners). If these tasks look ok I plan to take ATLAS out of beta. |
Send message Joined: 13 May 14 Posts: 387 Credit: 15,314,184 RAC: 0 |
New tasks are available now, but they look like they will run a bit longer than the previous batch. They are not as long as the "longrunners" but probably 4 times as long as before. |
Send message Joined: 2 May 07 Posts: 2071 Credit: 156,193,206 RAC: 103,326 |
2 CPU's and Virtualbox 5.1.18. Ended after 15 Minutes. https://lhcathome.cern.ch/lhcathome/result.php?resultid=127234256 Any idea? |
Send message Joined: 2 Sep 04 Posts: 453 Credit: 193,369,412 RAC: 10,065 |
2 CPU's and Virtualbox 5.1.18. Ended after 15 Minutes. Not shure if my clients have already picked up the latest WUs, but from my latest WUs the runtime has already climbed up to 50 minutes Supporting BOINC, a great concept ! |
Send message Joined: 14 Jan 10 Posts: 1268 Credit: 8,433,416 RAC: 3,056 |
2 CPU's and Virtualbox 5.1.18. Ended after 15 Minutes. From your stderr output: 2017-03-20 16:27:28 (5920): Setting Memory Size for VM. (3600MB) 2017-03-20 16:27:28 (5920): Setting CPU Count for VM. (2) As written before, a multi core (so also a dual core) needs at least 4400MB of RAM. You can manage that with an app_config.xml or select in your preferences 1 job with 3 cores to get a VM with 4600MB of RAM running. |
Send message Joined: 2 May 07 Posts: 2071 Credit: 156,193,206 RAC: 103,326 |
Thanks Crystal, the long-runner had 3 CPU's and finished correct. Atlas@home tasks with 2 CPU's had running without app_config.xml a long time for me. Than must Atlas under LHCatHome need more RAM? Had made no change in boinc-manager. |
Send message Joined: 14 Jan 10 Posts: 1268 Credit: 8,433,416 RAC: 3,056 |
Than must Atlas under LHCatHome need more RAM? Only for a dual core. It's not configured well for 2 core VM's, cause the server is using the formula 1.6GB + 1GB/core. For 1 core it's OK and for 3 and more cores also. |
Send message Joined: 18 Dec 16 Posts: 123 Credit: 37,495,365 RAC: 0 |
See this post for a bit more explanation on the changes between ATLAS and LHC: https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4166&postid=29459#29459 We are the product of random evolution. |
©2024 CERN