Message boards : ATLAS application : Small number of test tasks
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · Next

AuthorMessage
David Cameron
Project administrator
Project developer
Project scientist

Send message
Joined: 13 May 14
Posts: 387
Credit: 15,314,184
RAC: 0
Message 28473 - Posted: 13 Jan 2017, 16:06:39 UTC

We have started the integration of ATLAS into LHC@Home and we are sending a few test WU just to see if they work ok. These WU are exactly the same as the multicore WU on ATLAS@Home but we are not using the results. The ATLAS app is currently a beta app so you need to enable running beta apps in the project preferences if you want to try ATLAS WU. We welcome volunteers (especially seasoned ATLAS@Home crunchers!) to try out the test tasks.
ID: 28473 · Report as offensive     Reply Quote
Luigi R.
Avatar

Send message
Joined: 7 Feb 14
Posts: 99
Credit: 5,180,005
RAC: 0
Message 28474 - Posted: 13 Jan 2017, 16:17:21 UTC

Ok, I will see if I could run some tasks. Do you need a feedback? Only negative ones?
ID: 28474 · Report as offensive     Reply Quote
Luigi R.
Avatar

Send message
Joined: 7 Feb 14
Posts: 99
Credit: 5,180,005
RAC: 0
Message 28480 - Posted: 13 Jan 2017, 21:19:43 UTC
Last modified: 13 Jan 2017, 21:20:12 UTC

ID: 28480 · Report as offensive     Reply Quote
Claus Varming Lund

Send message
Joined: 15 Feb 15
Posts: 7
Credit: 26,671,002
RAC: 0
Message 28488 - Posted: 14 Jan 2017, 9:55:14 UTC

Hi
Same as Luigi, Validate error.
I ran on a Win10, with Boinc 7.6.33 and VB 5.1.12
Let us know when there are more ATLAS to test :-)
ID: 28488 · Report as offensive     Reply Quote
David Cameron
Project administrator
Project developer
Project scientist

Send message
Joined: 13 May 14
Posts: 387
Credit: 15,314,184
RAC: 0
Message 28521 - Posted: 16 Jan 2017, 13:02:54 UTC

I made a small mistake in one of the parameters of the WU, but I fixed that this morning and the new WU are failing with a different problem. I am investigating...
ID: 28521 · Report as offensive     Reply Quote
David Cameron
Project administrator
Project developer
Project scientist

Send message
Joined: 13 May 14
Posts: 387
Credit: 15,314,184
RAC: 0
Message 28534 - Posted: 17 Jan 2017, 11:18:37 UTC

After fixing some backend infrastructure issues, we now have the first successful WU! We will keep a small trickle of test WU in the system so let us know of any problems you see.
ID: 28534 · Report as offensive     Reply Quote
Pasi Nevalainen

Send message
Joined: 24 Jan 15
Posts: 2
Credit: 1,628,525
RAC: 0
Message 28536 - Posted: 17 Jan 2017, 18:26:17 UTC - in response to Message 28534.  

One task calculated now. All works fine :)
ID: 28536 · Report as offensive     Reply Quote
Profile Yeti
Volunteer moderator
Avatar

Send message
Joined: 2 Sep 04
Posts: 453
Credit: 193,369,412
RAC: 10,065
Message 28537 - Posted: 17 Jan 2017, 19:29:18 UTC

Will the Atlas-WUs here take into account the settings regarding NumberOfWUs and NumberOfCores from the Website or do we have to setup an app_config ?


Supporting BOINC, a great concept !
ID: 28537 · Report as offensive     Reply Quote
Profile Magic Quantum Mechanic
Avatar

Send message
Joined: 24 Oct 04
Posts: 1114
Credit: 49,501,728
RAC: 4,157
Message 28539 - Posted: 17 Jan 2017, 23:52:45 UTC

I tried to get some of these tasks but I see the server says 0 tasks to send and 12 in progress........of course I tried to get some first with no luck.
Volunteer Mad Scientist For Life
ID: 28539 · Report as offensive     Reply Quote
Profile Yeti
Volunteer moderator
Avatar

Send message
Joined: 2 Sep 04
Posts: 453
Credit: 193,369,412
RAC: 10,065
Message 28540 - Posted: 18 Jan 2017, 5:50:16 UTC - in response to Message 28473.  

These WU are exactly the same as the multicore WU on ATLAS@Home


I was lucky and got my first WU here.

If they are really the same as on Original-project, the rutime is to small. Please check this result: https://lhcathome.cern.ch/lhcathome/result.php?resultid=112410796

CPU-time as 4-Core-WU at Original-Atlas is for this machine something with 4 hours, here it is only something with 33 minutes, but the WU got validated.


Supporting BOINC, a great concept !
ID: 28540 · Report as offensive     Reply Quote
David Cameron
Project administrator
Project developer
Project scientist

Send message
Joined: 13 May 14
Posts: 387
Credit: 15,314,184
RAC: 0
Message 28541 - Posted: 18 Jan 2017, 8:22:05 UTC - in response to Message 28540.  

These WU are exactly the same as the multicore WU on ATLAS@Home


I was lucky and got my first WU here.

If they are really the same as on Original-project, the rutime is to small. Please check this result: https://lhcathome.cern.ch/lhcathome/result.php?resultid=112410796

CPU-time as 4-Core-WU at Original-Atlas is for this machine something with 4 hours, here it is only something with 33 minutes, but the WU got validated.


Sorry, I forgot to mention that these WU process 5 events instead of the normal 50 events used in the original project. This is to give a quick turnaround of tasks so problems can be seen quickly. So a short CPU time is expected.
ID: 28541 · Report as offensive     Reply Quote
Profile Laurence
Project administrator
Project developer

Send message
Joined: 20 Jun 14
Posts: 372
Credit: 238,712
RAC: 0
Message 28542 - Posted: 18 Jan 2017, 8:25:30 UTC - in response to Message 28537.  
Last modified: 18 Jan 2017, 8:56:40 UTC

Will the Atlas-WUs here take into account the settings regarding NumberOfWUs and NumberOfCores from the Website or do we have to setup an app_config ?


They should do as they act upon the scheduling in the server rather than the client. I am just testing now for myself but there are no tasks.

Edit: It doesn't work at the moment but we should be able to get it working.
ID: 28542 · Report as offensive     Reply Quote
Crystal Pellet
Volunteer moderator
Volunteer tester

Send message
Joined: 14 Jan 10
Posts: 1268
Credit: 8,421,616
RAC: 2,139
Message 28543 - Posted: 18 Jan 2017, 8:48:13 UTC
Last modified: 18 Jan 2017, 9:18:58 UTC

Last 3 tasks ended into an error

First showing the Login prompt and after a short time: "Can't connect to default. Skipping." ... and the VM shuts down.



https://lhcathome.cern.ch/lhcathome/result.php?resultid=112535871
https://lhcathome.cern.ch/lhcathome/result.php?resultid=112535831
https://lhcathome.cern.ch/lhcathome/result.php?resultid=112535833
ID: 28543 · Report as offensive     Reply Quote
Crystal Pellet
Volunteer moderator
Volunteer tester

Send message
Joined: 14 Jan 10
Posts: 1268
Credit: 8,421,616
RAC: 2,139
Message 28544 - Posted: 18 Jan 2017, 9:23:00 UTC - in response to Message 28543.  

Last 3 tasks ended into an error

No panic, it's working now.
The line <enable_shared_directory/> was missing.

btw: It would be nice to have at least Alt-F3 (top) enabled.
ID: 28544 · Report as offensive     Reply Quote
Crystal Pellet
Volunteer moderator
Volunteer tester

Send message
Joined: 14 Jan 10
Posts: 1268
Credit: 8,421,616
RAC: 2,139
Message 28545 - Posted: 18 Jan 2017, 10:54:33 UTC

Results of one single core VM and a dual_core VM:

https://lhcathome.cern.ch/lhcathome/result.php?resultid=112535885 --> Run time 55 min 5 sec - CPU time 49 min 52 sec
https://lhcathome.cern.ch/lhcathome/result.php?resultid=112552204 --> Run time 36 min 22 sec - CPU time 56 min 22 sec

It seems credit is assigned to elapsed time and not to used cpu-time.
ID: 28545 · Report as offensive     Reply Quote
Toby Broom
Volunteer moderator

Send message
Joined: 27 Sep 08
Posts: 798
Credit: 644,678,398
RAC: 235,495
Message 28546 - Posted: 18 Jan 2017, 19:43:27 UTC
Last modified: 18 Jan 2017, 19:47:20 UTC

I have had 20 total, there is 4 waiting to run. The 1st one on the 16th had validate error, the rest are good it seems.

They are all mcore on my PC.

I have all mine set to no limit which is 1core, I can check if goes up.
ID: 28546 · Report as offensive     Reply Quote
gyllic

Send message
Joined: 9 Dec 14
Posts: 202
Credit: 2,533,875
RAC: 0
Message 28547 - Posted: 18 Jan 2017, 20:36:55 UTC

I crunched 8 wu's today, which all worked fine (except for one, but that was my fault playing around with the ram settings).

is it possible that there are multicore wu's which are using the allocated cores (2 cores + 2 HT = 4 in my case) and multicore wu's which are just using 1 core, regardless how much cores are allowed by the boinc manager? because i havent changed my settings according to allowed cores for boinc and some wu used 4 and some used just 1 (or is it because i played with the ram settings).

Boinc-Manager: 7.4.22
OS: Linux 4.8
Virtualbox. 5.1.14
ID: 28547 · Report as offensive     Reply Quote
gyllic

Send message
Joined: 9 Dec 14
Posts: 202
Credit: 2,533,875
RAC: 0
Message 28548 - Posted: 18 Jan 2017, 21:49:04 UTC
Last modified: 18 Jan 2017, 21:50:17 UTC

ok, i crowed too soon. last two tasks failed with validate error:

https://lhcathome.cern.ch/lhcathome/result.php?resultid=112760511
https://lhcathome.cern.ch/lhcathome/result.php?resultid=112760579

unfortunately, the editing function in this forum only works until 60 mins after the post, so i had to make a new one...
ID: 28548 · Report as offensive     Reply Quote
David Cameron
Project administrator
Project developer
Project scientist

Send message
Joined: 13 May 14
Posts: 387
Credit: 15,314,184
RAC: 0
Message 28552 - Posted: 19 Jan 2017, 8:50:42 UTC - in response to Message 28548.  

Those two were from the first batch of bad WU which were still hanging around.

Thanks everyone for the testing! I've submitted some longer WU now, processing 50 events like in the regular ATLAS@home.

The #CPU settings are not working as expected, but Laurence is looking into it.

As for credit, see the discussion here which shows that running 1 core gives much more credit due to longer elapsed time. This seems to be a feature of the standard CreditNew algorithm that we are all using.
ID: 28552 · Report as offensive     Reply Quote
Toby Broom
Volunteer moderator

Send message
Joined: 27 Sep 08
Posts: 798
Credit: 644,678,398
RAC: 235,495
Message 28571 - Posted: 19 Jan 2017, 19:36:20 UTC
Last modified: 19 Jan 2017, 20:29:03 UTC

Scrub that comment about multicore, they were the actual atlas tasks.
ID: 28571 · Report as offensive     Reply Quote
1 · 2 · 3 · Next

Message boards : ATLAS application : Small number of test tasks


©2024 CERN