climateprediction.net (CPDN) home page
Thread 'TASKS IN PROGRESS DROPPING'

Thread 'TASKS IN PROGRESS DROPPING'

Message boards : Number crunching : TASKS IN PROGRESS DROPPING
Message board moderation

To post messages, you must log in.

AuthorMessage
ProfileJIM

Send message
Joined: 31 Dec 07
Posts: 1152
Credit: 22,363,583
RAC: 5,022
Message 47604 - Posted: 19 Nov 2013, 3:42:30 UTC

I noted that the �Tasks in Progress� indicator on the Server Status page now stands at 49,664. This is the first time that I ever remember it being below 50,000. The count has been dropping at the rate of about 350 per day.

I wonder how many of those are really running, and how many of them are WU�s that were lost in a way that server never found out about it and just haven�t timed out yet. Either way we are going to need a large influx of work soon or there are going to be an awful lot of empty cores. (HINT, HINT)

ID: 47604 · Report as offensive     Reply Quote
Steinar1965

Send message
Joined: 4 Sep 06
Posts: 79
Credit: 5,583,517
RAC: 0
Message 47610 - Posted: 19 Nov 2013, 9:22:47 UTC
Last modified: 19 Nov 2013, 9:23:02 UTC

I understand that you dont know exactly when there will be new wu's but pls give us some info so we feel we know if it is this week or after christmas
ID: 47610 · Report as offensive     Reply Quote
ProfileDave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4540
Credit: 19,039,635
RAC: 18,944
Message 47611 - Posted: 19 Nov 2013, 9:41:09 UTC

There has been talk in another thread about this month being when the BOINC server code is upgraded. It has also been made clear several times that the work depends on what the researchers want and them submitting this to the people at Oxford to generate the work units. I suspect that most of the time the researchers who pay Oxford uni don't give a lot of advance notice and there is not a lot anyone there can do about it.
ID: 47611 · Report as offensive     Reply Quote
Eirik Redd

Send message
Joined: 31 Aug 04
Posts: 391
Credit: 219,896,461
RAC: 649
Message 47612 - Posted: 19 Nov 2013, 9:58:46 UTC - in response to Message 47610.  
Last modified: 19 Nov 2013, 10:06:11 UTC

Rain on the parade, sorry, but if that 49K "tasks in progress" has to get down to zero before before the team upgrades the server software, yeah, it will be 50000/400 = several months.
No info is available on future work that might be available, not on this board.
Have seen some mentions last 6 months might be some new regional projects (SA,NZ) in beta almost ready to go.

But -- my advice -- run other projects but stay attached, because when there is a new batch or batches - the work will be demandingly long and require a big commitment to running month-long wu-s.

Me - the work I got now -- mostly re-issues of wu-s that first issued in May-June and the site re-issued Oct-Nov this year. Wu-s that failed 3-4 times.

The time-scale here at CPDN is so enormously slower than at most other BOINC sites. It is like a time-warp. With the long time to run the models, feed back results to the actual researchers and them to submit new batches -- but that's the nature of climate research and the long, and sometimes fragile models.

Not to mention the hardware and software problems. Those happen everywhere on BOINC, but likewise much slower (reasonably so, especially considering the result storage requirements here. Again - because the long-running models means that getting results that can be summarized means huge datasets timewise and spacewise - yeah-
frog in the teakettle - slow. Worth my machines and attention, to my mind.
ID: 47612 · Report as offensive     Reply Quote
ProfileDave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4540
Credit: 19,039,635
RAC: 18,944
Message 47614 - Posted: 19 Nov 2013, 11:22:41 UTC

Rain on the parade, sorry, but if that 49K "tasks in progress" has to get down to zero before before the team upgrades the server software, yeah, it will be 50000/400 = several months.


Nothing I have read here suggests that all the tasks have to be finished and cleared before the upgrade can take place, just that the programmers' time in implementing the upgrade will stop new batches of work coming on line because they won't have the time to do both things at once.
ID: 47614 · Report as offensive     Reply Quote
Eirik Redd

Send message
Joined: 31 Aug 04
Posts: 391
Credit: 219,896,461
RAC: 649
Message 47615 - Posted: 19 Nov 2013, 11:54:45 UTC - in response to Message 47614.  
Last modified: 19 Nov 2013, 12:00:13 UTC

Nothing I've read here suggests that it will be that long.
But nothing I've read here suggests that it will be sooner.
Just guessing.

Project policy seems to be - "make no commitments for future work"

Seems reasonable policy, given the multiple dependencies the project has to work with.

So, we're all just guessing. OK by me. Keep on crunching - that's my policy.

[edit]

And be prepared if/when a new bunch of work hits the queue. Could be a big bunch. I vacuum my heat sinks every month. :)
ID: 47615 · Report as offensive     Reply Quote
ProfileDave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4540
Credit: 19,039,635
RAC: 18,944
Message 47713 - Posted: 3 Dec 2013, 20:26:07 UTC

A whole bunch of new units have appeared. -:)
ID: 47713 · Report as offensive     Reply Quote
Albert H.

Send message
Joined: 18 Feb 06
Posts: 73
Credit: 61,754,697
RAC: 46,486
Message 47714 - Posted: 3 Dec 2013, 21:48:11 UTC

O Happy Day we can go on
ID: 47714 · Report as offensive     Reply Quote
ProfileastroWX
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 1496
Credit: 95,522,203
RAC: 0
Message 47717 - Posted: 4 Dec 2013, 0:01:58 UTC
Last modified: 4 Dec 2013, 1:13:11 UTC

Not many running yet but one failed immediately. Others started okay:
<stderr_txt>

Model crashed: INITTIME: Atmosphere basis time mismatch tmp/xaakm.pipe_dummy 2048
Leaving CPDN_Main::Monitor...
Called boinc_finish

</stderr_txt>

http://climateapps2.oerc.ox.ac.uk/cpdnboinc/results.php?hostid=1224202

[EDIT]Apologies. This should have been placed in a new thread.
"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.
ID: 47717 · Report as offensive     Reply Quote
ProfileBonsai911

Send message
Joined: 9 Sep 04
Posts: 228
Credit: 30,756,611
RAC: 3,303
Message 47718 - Posted: 4 Dec 2013, 5:09:00 UTC

Good news from Hamburg!

All WU's are fine.

Bonsai911
ID: 47718 · Report as offensive     Reply Quote
ProfileastroWX
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 1496
Credit: 95,522,203
RAC: 0
Message 47719 - Posted: 4 Dec 2013, 6:24:27 UTC

Good to see.

I hope I've soaked-up most of the bad ones (but doubt it). Found ten more in queues on my machines. Seems strange to have the issue with only part of the release.

Andy should be at work in a couple hours...

Meanwhile, I'm off to bed.
"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.
ID: 47719 · Report as offensive     Reply Quote

Message boards : Number crunching : TASKS IN PROGRESS DROPPING

©2024 cpdn.org