Just finished uploading 2 completed tasks (they happened to finish at the same time - hate it when that happens ;-)
But now not getting any new tasks "No tasks are available..." even though Server Status shows a couple hundred long runs ready to send. I temporarily changed my prefs to get short runs, but not getting any of those either.

What's up??

Thanks,

Win

Edited to add: the number of available tasks to send is slowly increasing...

Just finished uploading 2 completed tasks (they happened to finish at the same time - hate it when that happens ;-)
But now not getting any new tasks "No tasks are available..." even though Server Status shows a couple hundred long runs ready to send. I temporarily changed my prefs to get short runs, but not getting any of those either.

What's up??

Thanks,

Win

Edited to add: the number of available tasks to send is slowly increasing...

I too, have not been able to get a new WU. I have only one instead of two and when I try requesting one it says "No tasks available." The tasks ready to send keeps increasing. 3 of my computers are idling now.

I have two idle machines now with the same problem. If they don't send them, we can't do them.

It looks to me that only my Windows 10 computers are affected by this error, as my Windows XP computers received tasks since then.
Perhaps these new workunits have not been assigned to the CUDA 8.0 (v9.18) app.

Perhaps these new workunits have not been assigned to the CUDA 8.0 (v9.18) app.

Both of my machines are Ubuntu 16.04.3 using the v9.14 app on a GTX 980 and GTX 1070, both with the 387.12 (CUDA 9.1) driver. They have been working fine for weeks, then everything just stopped. A manual update does not work either.

So you are fortunate that it works on XP at least.

me too!
a Windows 7 and a Linux 17.04 box asking for work, not getting any.

When I wrote my post there were 600 unsent tasks. Now up to 1532. And tasks in progress has dropped by 1000.

Win

same problem with my Windows10 machine.

Most probably it now depends on whether or not the GPUGRID people are reading here about this problem and will set corrective action.

Most probably it now depends on whether or not the GPUGRID people are reading here about this problem and will set corrective action.

one thing which they should notice earlier than reading these postings here, is the fact that the number of unsent tasks is climbing up rapidly hour by hour.
Right now, it's close to 1.500 - a value which has never occurred in the recent past.

Further, the number of "task in progress" is falling hour by hour.

I may have made a mistake with the scheduler. Please check if it is back to normal.

I may have made a mistake with the scheduler. Please check if it is back to normal.

Now it's ok, thanks Toni !

---

Dreams do not always come true. But not because they are too big or impossible. Why did we stop believing.
(Martin Luther King)

Please check if it is back to normal.

yes, it is.
Many thanks, Toni

Sorry about that and thanks for noting that it is a cuda-dependent thing. I was changing the scheduler to enable sending out experimental "quantum chemistry" workunits (CPU).

Stuff happens.

Thanks for the fix!

I used the downtime to finally do a full nvidia driver uninstall (DDU tool) and upgrade to the latest driver. Previous upgrades were crashing on GPU tasks.

Win

I attempted to reintroduce the change. Please report if task distribution goes wrong.

I attempted to reintroduce the change. Please report if task distribution goes wrong.

Hi Tony,
task distribution works well, new Pablo long WU in process here.
K.

---

Dreams do not always come true. But not because they are too big or impossible. Why did we stop believing.
(Martin Luther King)