Boinc keeps telling me this project has no CPU or GPU WUs available for download

Hi David,

Check the "Server Status" page, on the blue bar to the left.

If there is work available (for HadAM3-EU), check your CPDN preferences, which you will find on the "Your account" page in the blue bar to the left. Make sure that you're allowing HadAM3-EU work ... but only if your machine can process it. Older AMDs and Pentium IIIs can't.

All projects are enabled and I'm running on Windows 7X64 with a i7-2600K

nm, i Just told it to allow new work again and it pulled down a bunch of WUs :)

The servers say that work is available but it won't give me any.

I have not (to my knowledge) set any preferences that would avoid any CP work. Anyone know what could be wrong?

Bill H

Check any messages on BOINC - UNDER Advanced - Event Log on the most recent versions. That may give a clue, also go to your account and make sure you are allowing the Hadcm3n tasks that are all that are available at the moment - Every so often a regional model pops up but unless your computer asks for it at exactly the right time you won't get it. Because the hadcm3n models have bigger files associated with them, if you don't have enough disk space allocated to BOINC it won't download them. (I recently discovered this on Linux but it will also apply to Windows.)

Dave

Just checked: I have all the projects in the list selected. I also have the box checked that allows substitution.
I have 8gb of RAM (some of which will be in use) and 500Gb of free disk (on the c: drive) and 4 AMD processors.
I am running BOINC 6.12.34 (64)
The response from the server was:-
30/10/2011 14:43:16 | climateprediction.net | Sending scheduler request: To fetch work.
30/10/2011 14:43:16 | climateprediction.net | Requesting new tasks for NVIDIA GPU
30/10/2011 14:43:18 | climateprediction.net | Scheduler request completed: got 0 new tasks
30/10/2011 14:43:18 | climateprediction.net | No work sent

(I think - couldn't look in the two places simultaniously) that there were 2 -EU jobs 'available'
Does this mean I'm only requesting GPU work? (and if so how do I change it?)

Bill H

Not sure about it only requesting GPU work - none of the CPDN models use the GPU, I do know that. All I can suggest is looking at the preferences and BOINC settings on your two computers, going through them line by line to see if the one not getting any work is different. I see one is running 4 tasks at present.

Dave

Bill

Some things that can inhibit work requests/downloads:
1) Having any work from the project of interest Suspended. (In this case, cpdn climate models.)
2) Having enough work already on hand for the time being.
3) Having a low resource share for the project.
4) Not running the computer for long enough periods. (Or YOU using the computer intensively.)
5) Having run so many long climate models, that your other projects are owed heaps of crunch time to keep the resource share balanced.
6) Having a small cache.

If the messages that you posted where in response to you clicking on Update, then they don't count. It's only if BOINC does the asking by itself and doesn't get any work, that matters.

If you really need more work, then the quickest fix may be to set the Computer is connected to the Internet about every to 10.0, and
Maintain enough work for an additional to say, 2.0

Note that if the cause for 'no work' is due to one of the other reasons, and BOINC suddenly decides that it's time for more work by itself, then increasing the cache size may result in you getting far too much work. This will interfere with your work fetch from other projects.

And if you DO end up with far too many models, then DON'T Abort them!
This won't return them to allow others to run them.
On the contrary, depending on the settings for the work units, it may well 'sterilize' the lot, so that there are NO re-issues.
Just Suspend them for later processing.

*****************

As for asking for GPU work, that's how BOINC finds out if a project has added a GPU app since the last time it asked.
It's possible to inhibit this asking on a per project basis, but not yet at cpdn.

The other two options are
1) A cc_config setting
2) A local prefs setting for a given computer

Both of the above prevent ANY GPU work from being done for ANY project.

---

Backups: Here

I see that there are presently no WUs available for either the Hadam3p or Hadcm3n projects. I would be nice if one of the research teams would throw a few thousand WUs into the hopper in the next couple of days.

---

Yes indeed.
But the research teams are situated at various places around the world, all with there own agendas.
So it's a matter of waiting. :)

---

Backups: Here

Great to see lots of work available now. At one point it looked as if there might be more tasks to send out than hosts attached to the project! First time I have seen that for a while!