Hi-

For a long time I have run climate prediction tasks but for the last month or so I don't have any. I checked the status and it seems the servers are up and running.

Any idea of why not tasks are being downloaded to run on my Mac (Mac Pro, Lion 10.7.2)?

Here is the response I get when I send and update to the project via BOINC:

Tue Jan 31 19:23:23 2012 | climateprediction.net | update requested by user
Tue Jan 31 19:23:25 2012 | climateprediction.net | Sending scheduler request: Requested by user.
Tue Jan 31 19:23:25 2012 | climateprediction.net | Not reporting or requesting tasks
Tue Jan 31 19:23:28 2012 | climateprediction.net | Scheduler request completed

Thanks,
Don

Your computer isn't asking for climate work, so it doesn't need any.
You're running several projects, so perhaps BOINC is in 'pay back time' after running a lot of the long climate models.

And you seem to be crashing a lot of them. Try only running the shorter regional models while there's plenty of them.

---

Backups: Here

Thanks for your quick reply Les. I have no clue why they were crashing, I guess I have been a bad person. 8-) Seriously I have been running climate models for a couple/three years now with no problems until the last year or so. Any ideas why there3 are crashes?

How can I tell if a climate model is a shorter regional model or not? Do I have control over what kind of models are downloaded for my computer to work on?

In the blue list on the left, go to Your account/Preferences for this project/scroll to your computer assignment (Default/Home/School/or Work), go to the bottom of that block and click Edit preferences. There is a selection list. Click to add or subtract task types; Regional models are PNW/EU/SAF. Click Update preferences at the bottom of the page. (This is a server update, so no necessity for updating boinc, as with other changes.)

Please read the boinc 6.12.26 Sticky at the top of the MAC section of the Board.

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Thanks, I selected only regional models to run now. We'll see how this goes.

FYI, I had already read and followed the 6.12.26 Sticky a while back, first the "reset project" and then the "detach/attach". Neither seemed to have worked.

Don,

Another thing to try is to reduce the number of times the models are suspended. A number of the errored tasks on your machine have numerous suspends, which can cause problems - particularly with the longer HADCM3N tasks. The most likely reason for the large number of suspends is a default BOINC setting in the 'processor usage' section of the BOINC Manager preferences. Have a look at the entry that starts 'while processor usage is less than': the default is 25% - setting it to 0% stops BOINC suspending so often. The default setting is intended to make BOINC as unobtrusive as possible, but for most people that protection is more trouble than it's worth. (The tasks still run at low priority, so shouldn't interfere with normal activity.)

Iain

Since I have upgraded to 7.0.25 on my Mac, none of my projects fetch WUs automatically as they used to do. My preferences are the same, ie Max work buffer set to 2 days, there are only a few projects currently running with a small number of WUs that I had to fetch by Updating manually. Any ideas?

Since I have upgraded to 7.0.25 on my Mac, none of my projects fetch WUs automatically as they used to do. My preferences are the same, ie Max work buffer set to 2 days, there are only a few projects currently running with a small number of WUs that I had to fetch by Updating manually. Any ideas?

7.0 has some major changes to the way the BOINC client operates which haven't been widely publicised.

1. Work scheduling
   
   The concept of project debts (short and long term) are no longer used. The work scheduler now uses something called Recent Estimated Credit (REC). This measures how long your system takes to run work for each project, accumulating when a project is running and decaying when it isn't running. It isn't directly related to RAC or actual credit (which are controlled by the project servers).
   
   
2. Work fetch
   
   Work fetch is now dependent on the REC and resource share ratio for each project. The project with the lowest REC will, in general, have the highest priority when requesting work.
   
   
3. Global preferences
   
   The interpretation of the first 2 parameters in the network usage section of global preferences has changed in BOINC 7.0
   
   
   
   
   • Computer is connected to the Internet about every is now displayed on the Network Usage tab of the Preferences dialog (accessible from the Avanced - Preferences menu item in BOINC Manager's advanced view) as Minimum work buffer with the tooltip Try to maintain enough tasks to keep busy for this many days. I'll refer to this as minimum below.
     
     
   • Maintain enough work for an additional is now displayed on the same tab as Max additional work buffer with tooltip In addition, maintain enough tasks for up to this many days. I'll refer to this as additional below.
     

   
   When the total amount of work for all projects on your system for a resource (e.g. CPU or GPU) falls below "minimum" BOINC will request enough work from the highest priority project to top the amount of work up to "minimum" + "additional".
   
   If both values are set to 0 BOINC won't request work until you have an idle CPU (or GPU).
   

One consequence of these changes is that computers are more likely to have periods when one or more projects have no pending work.

---

"The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer

Thanks for the help. I had never set the minimum work buffer before, it was always zero. Now I have set it to 1 day and the max is set to 2 days. It works fine now. Thanks again!