Hi everyone,
Two members of my team have a problem in the simultaneous processing of their tasks SETI@home (without differentiation of the nature of the tasks - CPU, GPU or CUDA).
Their computers have 8 processors and NVIDIA GTX. However, BOINC Manager (version 7.0.28) processes only one task at a time.
We checked all the configuration processing tasks but nothing has changed from what they originally set. Dysfunction has emerged recently.
Previously, they treated eight tasks simultaneously without ever having any problems.
Does anyone have an explanation or advice to give me about this ?

A small correction - the BOINC Manager doesn't process anything at all. It's simply there so that you can interact with, and control, the BOINC 'client' (worker) - which is otherwise invisible to you. You don't have to be running the Manager for the client to be working.

But that's a minor quibble.

What you do need to establish from your team-mates is: is there work available on their computers for the other resources (I'm guessing CPUs) to process? That is, work assigned, downloaded, and shown as "Ready to start"?

If there is no work, then it can't be processed - and you have to find why there is no work. What happens when they click the 'update' button? What is shown in the event log? If it isn't immediately clear from the log, you may have to enable additional debug logging - <sched_op_debug> is a good place to start, followed by <work_fetch_debug> if you still need it.

But the only way you can solve an issue like this is to gather, and read carefully, information from the machines in question. Nobody here can give you advice without that.

Richard,
Thanks very much for your quick response.
Here's a screenshot of the tab "Tasks" one of my guys.
We set a reserve of tasks and there has a packet waiting to start.
Everything seems normal.
http://img11.hostingpics.net/pics/785619Capture100.png
I remain at your disposal if you want more information.

All the tasks visible in that screenshot are scheduled to run on the (single) NVidia GPU - you can tell that from the 'Application' column, saying "6.10 (cuda_fermi)". Any tasks assigned to the (8x) CPUs would show as 6.03, with nothing in brackets. You need to scroll the list and look for those. If none are visible, then you are either not requesting, or not being assigned, CPU work.

If not requesting, check configuration settings, both in the Manager and on the website - verify using Event log that your changes are successful.

If requesting, but not receiving, that's a project issue. Read the (regular) new threads on the SETI project message boards about why this happens.

THX for your help Richard.

Hi all,

I'm having the exact same problem with only one task being worked at a time, the one using the NVIDIA GPU, even though I have 4 processors on my computer. I'm using v. 7.0.28 and everything was working fine until a few days ago. Now only the applications downloaded are "6.10 (cuda_fermi)". I've checked all my settings, both in the BOINC Manager and online, and there's nothing I see that limits what tasks to download. I also tried removing BOINC and reinstalling it. Same thing.

Is there anything else I can try or is this a problem with BOINC/SETI and I just have to wait it out?

Thanks.

What cache settings are you running? the server's scheduler will fill up the cache with work for fastest device first, ie the GPU, once that's full then it'll send work for the CPU,
you can fix this by lowering the cache settings until Boinc stops asking for GPU work,

Claggy