climateprediction.net home page
HADAM3P not getting credits

HADAM3P not getting credits

Message boards : Number crunching : HADAM3P not getting credits
Message board moderation

To post messages, you must log in.

AuthorMessage
Nigel Conway

Send message
Joined: 20 Dec 05
Posts: 14
Credit: 277,045
RAC: 0
Message 49139 - Posted: 17 May 2014, 17:39:35 UTC

I have been running climate CPDN for many years without significant issues. In April 5 HADAM3P Australia New Zealand v6.10 tasks were downloaded. Since then no credits have been allocated. One task has has failed and the other 4 have over 9 hours of elapsed time. I run other projects on the same machine and they are fine.
ID: 49139 · Report as offensive     Reply Quote
Profile Ray Murray
Avatar

Send message
Joined: 7 Aug 04
Posts: 50
Credit: 548,730
RAC: 0
Message 49140 - Posted: 17 May 2014, 19:25:42 UTC - in response to Message 49139.  

Hi Nigel,
The ANZ tasks are a bit slower than others, presumably because they are more complex. The model I started a couple of days ago on a 2.5 GHz machine took 22hours to get to the first trickle, 6.8 sec/Ts where other tasks have been around 2 secs/Ts. They also only have 12 trickles and it needs to get to a trickle point to have credits awarded. If the model looks like it running fine then it probably is, but after 9 hours, it isn't far enough along to have trickled yet, therefore no credits yet,
Patience is the key.
ID: 49140 · Report as offensive     Reply Quote
Nigel Conway

Send message
Joined: 20 Dec 05
Posts: 14
Credit: 277,045
RAC: 0
Message 49141 - Posted: 17 May 2014, 23:01:14 UTC - in response to Message 49140.  

Hi Ray,
Thanks for the confidence booster. It just seemed after a month and a half there should have been some credits. The most advanced is at 11:42:19 elapsed and 143:08:46 so should be almost there.

I'll give it a while longer.
ID: 49141 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 49147 - Posted: 18 May 2014, 7:44:27 UTC - in response to Message 49141.  

On my machines, a complete model takes about 110 hours, a bit over 4 days.

If yours are running that slowly, you're either not leaving the computer on for very long each day, you haven't given cpdn as much priority as your other projects, or there's something wrong with the computer.

ID: 49147 · Report as offensive     Reply Quote
Richard Haselgrove

Send message
Joined: 1 Jan 07
Posts: 925
Credit: 34,100,818
RAC: 11,270
Message 49149 - Posted: 18 May 2014, 9:52:46 UTC - in response to Message 49141.  

The task which failed on 3 May shows a huge number of "Suspend request from BOINC..." in stderr.txt

You might perhaps review your BOINC preferences, so that the climate models have a clear run without being constantly interrupted.
ID: 49149 · Report as offensive     Reply Quote
Nigel Conway

Send message
Joined: 20 Dec 05
Posts: 14
Credit: 277,045
RAC: 0
Message 49155 - Posted: 18 May 2014, 17:58:18 UTC - in response to Message 49147.  

The HADAM3P Australia New Zealand models seem to be absolutely huge. presumably that's why the deadline is 21st March 2015. This is running on a MAC which I also use for video editing so I need the resource hence the suspend events. The machine is turned off quite a bit as I also use other machines.

I'll just let the tasks run and see what happens.


ID: 49155 · Report as offensive     Reply Quote
Nigel Conway

Send message
Joined: 20 Dec 05
Posts: 14
Credit: 277,045
RAC: 0
Message 49210 - Posted: 25 May 2014, 1:19:41 UTC - in response to Message 49155.  

Hi! I now have a task over 9% complete and no credit. Seems that something is not working. Are there any logs I can check?

I also had a failed task with 23,375.33 cpu sec and no credit. It seems like there is a problem here. I'm not too fussed about my credit but it could put off some people!!!


ID: 49210 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 49211 - Posted: 25 May 2014, 1:57:58 UTC - in response to Message 49210.  

The log to check is the Events log, which will have a message every time there's a scheduler request to upload a trickle_up file, which is the start of the credits process.
Press: Control + shift + E to open this log.

You could also read through some/all of this thread, which is all about the credits issue, although it's starting to drift a little.
It's called: Credit updates?, and is just below your thread.


ID: 49211 · Report as offensive     Reply Quote
Nigel Conway

Send message
Joined: 20 Dec 05
Posts: 14
Credit: 277,045
RAC: 0
Message 49240 - Posted: 29 May 2014, 0:20:15 UTC - in response to Message 49211.  

Hi Les,
Event log does not mention trickles and is not much help. Is anyone else geting credit on this model or has someone forgotten to add trickles to it? :-)

Thanks for the thread reference but that seems to be about external stats sites. I'm looking at the stats from boinc manager and the cpdn website.I'm assuming they are pretty authoritative. I will just let these tasks run and see what happens.
ID: 49240 · Report as offensive     Reply Quote
Profile Greg van Paassen

Send message
Joined: 17 Nov 07
Posts: 142
Credit: 4,271,370
RAC: 0
Message 49241 - Posted: 29 May 2014, 1:11:49 UTC - in response to Message 49240.  
Last modified: 29 May 2014, 1:12:16 UTC

hi Nigel,

please copy and paste your event log to a pastebin and post the URL here - or in a private message to Les, if you think here is too public.

The next step turn on debugging in your cc_config.xml - searching this message board via google ("site:climateapps2.oerc.ox.ac.uk cc_config.xml debugging"))ought to provide you with several sets of instructions on how to do that.
ID: 49241 · Report as offensive     Reply Quote
Profile Alan K

Send message
Joined: 22 Feb 06
Posts: 484
Credit: 29,579,234
RAC: 4,572
Message 49243 - Posted: 29 May 2014, 8:28:00 UTC - in response to Message 49240.  

Trickles on the HADAM3P models appear to be less frequent than on the full coupled models. Typically first trickles on anz models are round about 8 - 9% completion rather than 2.5% for a full coupled model. As for credits.....
ID: 49243 · Report as offensive     Reply Quote
Richard Haselgrove

Send message
Joined: 1 Jan 07
Posts: 925
Credit: 34,100,818
RAC: 11,270
Message 49244 - Posted: 29 May 2014, 8:42:19 UTC - in response to Message 49243.  

Trickles on the HADAM3P models appear to be less frequent than on the full coupled models. Typically first trickles on anz models are round about 8 - 9% completion rather than 2.5% for a full coupled model.

Err, yes.

A full coupled model is simulating 40 years of weather. It reports in via a trickle once every year, or at each 2.5% increment.

An AM3P model is simulating one year of weather, and reports in via a trickle once every (simulated, idealised, 30-day) month. So, a trickle every 100/12, or 8.333% increment. Those are exact figures: no 'typical' or 'appears' about it.

If you care to dig a little deeper, you can see that for coupled models:
First trickle at timestep 25,920
Last trickle at timestep 1,036,800

and for AM3P (ANZ) models:
First trickle at timestep 11,819
Last trickle at timestep 138,539
ID: 49244 · Report as offensive     Reply Quote
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4314
Credit: 16,377,675
RAC: 3,657
Message 49245 - Posted: 29 May 2014, 9:04:38 UTC

If you are turning the computer off every day the log is certainly be default cleared and a blank one is started when BOINC starts up again. Don't know if this is relevant or not.
ID: 49245 · Report as offensive     Reply Quote
Richard Haselgrove

Send message
Joined: 1 Jan 07
Posts: 925
Credit: 34,100,818
RAC: 11,270
Message 49247 - Posted: 29 May 2014, 9:16:10 UTC - in response to Message 49245.  

If you are turning the computer off every day the log is certainly be default cleared and a blank one is started when BOINC starts up again. Don't know if this is relevant or not.

And logs for the previous session (and before that, and before that...) can be found in the files 'stdoutdae.txt' and 'stdoutdae.old' in the BOINC data directory.
ID: 49247 · Report as offensive     Reply Quote
Nigel Conway

Send message
Joined: 20 Dec 05
Posts: 14
Credit: 277,045
RAC: 0
Message 49363 - Posted: 15 Jun 2014, 0:24:22 UTC - in response to Message 49247.  

Task finished with computing error! No more tasks to run at present.
ID: 49363 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 49364 - Posted: 15 Jun 2014, 2:03:55 UTC - in response to Message 49363.  

ALL of your tasks are failing.
Possibly because you've left the setting: Suspend work if CPU usage is above
0 means no restriction
at the default of 25%.
Which is fine for other projects, but NOT here.

ID: 49364 · Report as offensive     Reply Quote

Message boards : Number crunching : HADAM3P not getting credits

©2024 climateprediction.net