climateprediction.net home page
VANISHING WU'S

VANISHING WU'S

Message boards : Number crunching : VANISHING WU'S
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 · Next

AuthorMessage
Eirik Redd

Send message
Joined: 31 Aug 04
Posts: 391
Credit: 219,888,554
RAC: 1,481,373
Message 47400 - Posted: 25 Oct 2013, 7:59:39 UTC

Actually there is more work, has been the last week.
Not enough to result in a backed up queue on the server status page -
But my machines are set for a two-day queue and haven't run out of work yet.

Big back-ed up queue on sever status page is not equal to "no new work"

Might take a day or two to get a new model - but they are there.
ID: 47400 · Report as offensive     Reply Quote
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4314
Credit: 16,378,503
RAC: 3,632
Message 47401 - Posted: 25 Oct 2013, 9:50:29 UTC - in response to Message 47400.  

I think it is luck of the draw. I have work buffer set for three days but since I fixed my fresh install of Ubuntu by adding the 32bit libs, i haven't had any work. Not sure but I think I trashed my ssd on netbook by crunching so either no crunching when I replace it or I go back to the old fashioned mechanical disk!
ID: 47401 · Report as offensive     Reply Quote
Eirik Redd

Send message
Joined: 31 Aug 04
Posts: 391
Credit: 219,888,554
RAC: 1,481,373
Message 47402 - Posted: 25 Oct 2013, 9:57:49 UTC - in response to Message 47401.  

Luck of the draw is the word.
No work is false.
Keep crunching!
ID: 47402 · Report as offensive     Reply Quote
ed2353

Send message
Joined: 15 Feb 06
Posts: 137
Credit: 33,347,857
RAC: 0
Message 47403 - Posted: 25 Oct 2013, 18:14:12 UTC - in response to Message 47400.  

My work buffer is as the default.
How do I increase it?
I assume you modify the

<work_buf_min_days>0.100000</work_buf_min_days>

<work_buf_additional_days>0.250000</work_buf_additional_days>

in Global Prefs, but to what values?
Thanks
ID: 47403 · Report as offensive     Reply Quote
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4314
Credit: 16,378,503
RAC: 3,632
Message 47405 - Posted: 26 Oct 2013, 8:21:09 UTC - in response to Message 47403.  
Last modified: 26 Oct 2013, 8:23:14 UTC

I ususally increase mine to the maximum allowed 10 days but as I am crunching WCG till I can get some CPDN work I have set it to three days at the moment. If using BOINC manager you can use tools>computing preferences>network preferences to do it through the GUI.
ID: 47405 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 47406 - Posted: 26 Oct 2013, 9:15:23 UTC

ed2353

Or it can be changed on the projects web site. There's a button in the Projects tab that takes you to your account page.

Note that for BOINC version 6.*, the values are 10 (days) + 2 or 3 (extra days), but for version 7.*, it's the other way around. The first value is the "high water mark", and the second figure is the "low water mark". And the work cache starts at "high", decreases to "low", then refills.

ID: 47406 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 47407 - Posted: 26 Oct 2013, 9:27:25 UTC

Eirik, Misty

There has been NO big batch of new work loaded to the queue, or announced, (privately or publicly).

If there IS work, then it's resends from failed/abandoned work, and/or auto-regens from returned, completed, models.

ID: 47407 · Report as offensive     Reply Quote
ed2353

Send message
Joined: 15 Feb 06
Posts: 137
Credit: 33,347,857
RAC: 0
Message 47408 - Posted: 26 Oct 2013, 12:41:18 UTC - in response to Message 47406.  

Thanks Les and Dave.
I don't know how I did not see those options when I looked originally.
I'll blame it on my age I guess!
ID: 47408 · Report as offensive     Reply Quote
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4314
Credit: 16,378,503
RAC: 3,632
Message 47411 - Posted: 27 Oct 2013, 7:39:01 UTC - in response to Message 47408.  

Interesting Les, I hadn't noticed the change when I went on to 7.x. Just checked however and I have filled it in the correct way around. :)
ID: 47411 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 47412 - Posted: 27 Oct 2013, 8:18:33 UTC - in response to Message 47411.  

Hi Dave

Yes, it's a bit obscure; depends on how much people read the BOINC/dev site messages, or the Release Notes, or the FAQ: BOINC 7.0.
In the Notes it's near the top under Scheduler changes; in the FAQ, near the bottom under In finer detail.

ID: 47412 · Report as offensive     Reply Quote
Profile JIM

Send message
Joined: 31 Dec 07
Posts: 1152
Credit: 22,053,321
RAC: 4,417
Message 47413 - Posted: 27 Oct 2013, 16:57:59 UTC - in response to Message 47407.  

Eirik, Misty

There has been NO big batch of new work loaded to the queue, or announced, (privately or publicly).

If there IS work, then it's resends from failed/abandoned work, and/or auto-regens from returned, completed, models.



Yes I�ve noticed that the WU�s that I have been getting lately all seem to have final numbers that are 3 or 4 indicating that these are reissues.

One thing that I have been wondering is, is it possible to set different work buffer settings on the same machine for different projects.

When I am running only CPDN (most of the time) I keep the work buffer set at 10 days so that the machines starts to beg for new work as early as possible. I feel that this helps keep all my cores full when work is scarce (which is most of the time now).

When I do have an empty core or 2 because CPDN has no work I allow it to run other projects. The problem is that I have to remember to change the work buffer setting (usually to 0.5 days. I don�t want to download 10 days worth of work from MaleriaControl or WCG!), then change it back to 10 days to troll for CPDN work.

Is it possible to set different work buffer settings for each project, such as by telling one project that the computer is at home and another project that the computer is at work?




ID: 47413 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 47414 - Posted: 27 Oct 2013, 19:18:51 UTC - in response to Message 47413.  

I don�t want to download 10 days worth of work from MaleriaControl ...
:)
Yes, been there!

And the solution IS, as you say, to use different venues.


ID: 47414 · Report as offensive     Reply Quote
Eirik Redd

Send message
Joined: 31 Aug 04
Posts: 391
Credit: 219,888,554
RAC: 1,481,373
Message 47436 - Posted: 29 Oct 2013, 9:39:48 UTC - in response to Message 47419.  

I've not extended buffer time, staying at 2 days - have several idle cores and want them ready when work arrives to get fast turnaround. (And doing statistics on per-core and ht-core productivity) Fill in a bit with other fave projects -give their tiny wu-s low buffer time.

Trying to keep my mini-server farm upgraded, stable, and ready for the next big batch to hit the fan, when and if :)

Keep on crunching.

e
ID: 47436 · Report as offensive     Reply Quote
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4314
Credit: 16,378,503
RAC: 3,632
Message 47591 - Posted: 17 Nov 2013, 15:32:00 UTC

Finally got a task, though it has a, "4" at the end of it. I preesume that means it hasn't completed for three other people. The others in the work unit have either errored out or timed out with no response. will see how it goes.
ID: 47591 · Report as offensive     Reply Quote
Profile astroWX
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 1496
Credit: 95,522,203
RAC: 0
Message 47592 - Posted: 17 Nov 2013, 19:57:47 UTC

Click on Task>Work Unit ID to see fate of earlier releases:
http://climateapps2.oerc.ox.ac.uk/cpdnboinc/results.php?hostid=1298411

Two were abandoned, one had Linux lib. issue, one was possibly computer and/or settings related.
"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.
ID: 47592 · Report as offensive     Reply Quote
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4314
Credit: 16,378,503
RAC: 3,632
Message 47968 - Posted: 14 Jan 2014, 12:22:52 UTC

Looks like another batch of 3CNs have been pulled. Only just over half way through mine so they would have gone by the time I was ready for more anyway.
ID: 47968 · Report as offensive     Reply Quote
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4314
Credit: 16,378,503
RAC: 3,632
Message 47970 - Posted: 14 Jan 2014, 12:46:19 UTC - in response to Message 47968.  

See the post in News and announcements.
ID: 47970 · Report as offensive     Reply Quote
Alex Plantema

Send message
Joined: 3 Sep 04
Posts: 126
Credit: 26,363,193
RAC: 0
Message 47971 - Posted: 14 Jan 2014, 14:35:24 UTC

I got three of these. I aborted them and now someone else is running them...
ID: 47971 · Report as offensive     Reply Quote
Profile JIM

Send message
Joined: 31 Dec 07
Posts: 1152
Credit: 22,053,321
RAC: 4,417
Message 47972 - Posted: 14 Jan 2014, 14:44:03 UTC

Just to confirm. WU�s that start Hadcm3n_7o0n, Hadcm3n_7kqc, Hadcm3n_79ze, Hadcm3n_7dj7 and Hadcm3n_7dj8 all for the year 1980 are the escapees and should be aborted?

ID: 47972 · Report as offensive     Reply Quote
Alex Plantema

Send message
Joined: 3 Sep 04
Posts: 126
Credit: 26,363,193
RAC: 0
Message 47973 - Posted: 14 Jan 2014, 14:49:30 UTC - in response to Message 47972.  

Right, but aborting them seems of little use as long as they are picked up by someone else immediately.
ID: 47973 · Report as offensive     Reply Quote
Previous · 1 · 2 · 3 · 4 · 5 · Next

Message boards : Number crunching : VANISHING WU'S

©2024 climateprediction.net