1st, my english is worse.
only easy answer formulation please.

I have a 2-core-AMD processor @2,2Ghz with 4GB 800Mhz and i thought that both cores are busy when I start this project. But my task-manger shows me that only my second core is busy.

What could be the problem?

You have to tell BOINC how many processors your computer has.

Two places:
1) On your Account page on this site, under Computing preferences, there are 2 choices:
A) For BOINC version 6, On multiprocessors, use at most
Enforced by version 6.1+ This is a percentage, e.g. 25%, 50%, 75%, or 100%
B) For earlier versions: On multiprocessors, use at most
Enforced by version 5.10 and earlier This is a number, e.g. 1, 2, 3, or 4

2nd place:
There is an over-ride option in your BOINC manager, in the menu under Advanced, Preferences. This is only for BOINC version 6.

---

Backups: Here

I remember at a other project, there was at my task-manager alwasy 2 tasks with 50%. now I have only one with 50%.

This are my settings
http://www.speedshare.org/download.php?id=3522712312
http://www.speedshare.org/download.php?id=D98F7A6D12

That are two jpegs

Open your BOINC Manager and do this please:

* Click on the Advanced menu.
* Click on Preferences (that\'s item #2 in the Advanced menu). A Preferences window will open.
* Select the Processor Usage tab (that\'s the first tab).
* It should say \'On multiprocessor systems, use at most 100% of the processors\'. If it doesn\'t say 100%, change the %. Das meint, alle CPUs dieses Computers benutzen.
* It should also say \'Use at most 50% of CPU time. If it doesn\'t say 50%, change the number.
* Click the OK button.
* You may now need to exit completely from BOINC and then restart BOINC.

---

Cpdn news

I have done all these things on my BOINC 6.4.7 - both on the client and in my web site settings.

I have one cpdn job with over 1,000 hours to run.

My other cpdn job finished.

Since then BOINC will not request another cpdn job ("requesting 0 tasks"), but if I connect to a second BOINC project, it immediately downloads jobs and starts running them on the second core.

Change my BOINC version maybe?

Patience will cure it.
Your BOINC is most likely in "pay back mode" to your other project(s) because of all of the time spent running cpdn.

BOINC doesn't run ALL of your projects ALL of the time, just on average. (Over say 6 months.)

---

Backups: Here

I normally ONLY run CPDN. In only attached to a second project after this problem started, to see what the behaviour would be. Even with all other projects disabled and with only 1 task (the single CPDN job it is running) it simply refuses to download additional CPDN work.

BOINC 6.2.19 on a different PC does not exhibit this behaviour.

UPDATE: Upgraded to BOINC 6.6.20 and it immediately did what I expected, i.e. realised it needed more work and downloaded the tasks.

25/04/2009 11:12:32 AM Version change (6.4.7 -> 6.6.20)
<snip>
25/04/2009 11:13:07 AM climateprediction.net Sending scheduler request: To fetch work.
25/04/2009 11:13:07 AM climateprediction.net Requesting new tasks
25/04/2009 11:13:12 AM climateprediction.net Scheduler request completed: got 2 new tasks
25/04/2009 11:13:14 AM climateprediction.net Started download of hadam3p_6.06_windows_intelx86.exe
25/04/2009 11:13:14 AM climateprediction.net Started download of hadam3p_data_6.06_windows_intelx86.zip

Wonder if my problem was related to the change to HADAM3P 6.06?

Hi Adrian

It looks as if it was indeed the BOINC upgrade that solved the problem. But the problem perhaps shouldn't have happened in the first place, and if it was caused by an unrealistic resource share for different projects on this computer, the upgrade should probably not have solved it.

However, the BOINC scheduler has become so much more complex recently that I've abandoned my attempts to understand it. As far as I can see, there may only be about half a dozen people in the world who now truly understand what it does (or in some cases what it should do).

The scheduler's overriding priority is supposed to be to finish tasks on time. I thought its next priority was to try to keep all cores busy while respecting the member's preferences.

Adrian, if this computer has work fetch problems again we may need to ask for extra information.

---

Cpdn news

At the moment, i use also BOINC Version 6.4.7. and i had the same problem. But because I crunch normally for two projects, I didn't care.

Maybe it is a specific problem with the Version 6.4.7.

After I'm going to do an upgrade, I'll tell.

Hope my englisch was readably ;)

Your English is very readable indeed.

Some recent BOINC alpha versions have had work fetch and scheduling problems but I think the BOINC scheduler has functioned generally well in public versions.

If a core is idle (not busy) with no work from any project, the scheduler may not be working correctly.

If the computer has processed a lot for CPDN, the other project(s) may need extra work to balance and respect the resource share you have decided for this computer. You may need to crunch exclusively for the other project for days or weeks to give each project its correct resource share. Members attached to two or more projects may not always have a climate model for every core.

If a task from a project is suspended, BOINC refuses to fetch more work from that project.

---

Cpdn news

So, a few days ago I did a BOINC-Upgrade 6.4.7 -> 6.6.20 (because of GPUGrid). Since the upgrade, there are 2 tasks on my 2-Core-Machine. At the moment both running because of long term debts.

My guess: That's really a special problem with BOINC version 6.4.7