12/28/2012 9:02:53 AM | GPUGRID | Sending scheduler request: To fetch work.
12/28/2012 9:02:53 AM | GPUGRID | Requesting new tasks for NVIDIA
12/28/2012 9:02:55 AM | GPUGRID | Scheduler request completed: got 0 new tasks
12/28/2012 9:02:55 AM | GPUGRID | No tasks sent
12/28/2012 9:02:55 AM | GPUGRID | No tasks are available for Long runs (8-12 hours on fastest card)
12/28/2012 9:02:55 AM | GPUGRID | No tasks are available for the applications you have selected.
12/28/2012 9:02:55 AM | GPUGRID | This computer has finished a daily quota of 1 tasks
12/28/2012 9:12:34 AM | GPUGRID | Sending scheduler request: To fetch work.
12/28/2012 9:12:34 AM | GPUGRID | Requesting new tasks for NVIDIA
12/28/2012 9:12:36 AM | GPUGRID | Scheduler request completed: got 0 new tasks
12/28/2012 9:12:36 AM | GPUGRID | No tasks sent
12/28/2012 9:12:36 AM | GPUGRID | No tasks are available for Long runs (8-12 hours on fastest card)
12/28/2012 9:12:36 AM | GPUGRID | No tasks are available for the applications you have selected.
12/28/2012 9:12:36 AM | GPUGRID | This computer has finished a daily quota of 1 tasks
12/28/2012 9:34:11 AM | GPUGRID | Sending scheduler request: To fetch work.
12/28/2012 9:34:11 AM | GPUGRID | Requesting new tasks for NVIDIA
12/28/2012 9:34:14 AM | GPUGRID | Scheduler request completed: got 0 new tasks
12/28/2012 9:34:14 AM | GPUGRID | No tasks sent
12/28/2012 9:34:14 AM | GPUGRID | No tasks are available for Long runs (8-12 hours on fastest card)
12/28/2012 9:34:14 AM | GPUGRID | No tasks are available for the applications you have selected.
12/28/2012 9:34:14 AM | GPUGRID | This computer has finished a daily quota of 1 tasks

12/28/2012 9:02:55 AM | GPUGRID | This computer has finished a daily quota of 1 tasks

That's why this host can't get any new tasks, but I don't know what, who or where set this quota to 1.

I uninstalled... rebooted... re installed vid drivers ... i CAN NOT figger this out :-( also, are there anymore Nathan tasks? They gave me the best credits, and i keep getting cuda 31 tasks (when i can get anything that is)Im running a 580 classified

A string of failures could cause this.

It's difficult to help people when their computers are hidden!

I suggest selecting to only run short tasks to improve your systems rating, though it will still take some time.

The 3.1 app is a pain that will go away early next year. If you are still getting lots of these try Zoltan's work around.

The most common causes of failures is heat and overclocking, so make sure you have your fans high enough to keep the GPU around 70°C or below. The more the GPU goes above 70°C the more likely it is that you will get failures. Over 80°C is just asking for trouble.

I'm just getting NOELIA_hfXA_long tasks at present. Given that 20,000 were put into the queue, it would be 10/1 at best to get another type of task. So if you keep aborting tasks your rating will plummet and the server will stop sending you work.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

i was at one time number 22 on the volunteers list then all of a sudden bam... just kept dropping because i was trying to get nathan tasks and aborting alot of 31's arrrrggggggg!!!! how can i get back to where i was? imm dropping like a rock. i havent changed a thing??? and im only overclocked a tiny bit.

Stop aborting tasks and just take what you can get. Other than a few resends there may only be NOELIA_hfXA tasks in the queue.

19 people now have a RAC of over 1M.
One GTX580 might just about get you back into the top 100 for RAC.
While it won't allow you to rise in the overall rankings to where you were (22nd) you should be able to get back into the overall top 200 in a few days and then slowly move up the ranks towards the top 100 for overall credit.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Thanks for the info... I have 3 puters ... running Only gpugrid 2 of them have 580 classified cards and one has a 660m, that 660m takes FOREVER to crunch numbers. VERY FRUSTRATING to be up at the top, then crap! I guess it was because i kept aborting the 31's ... i guess i will not do THAT anymore... thanks again and Merry Christmas :-)

I was in the top 10, and I'm now 27th. Even with 3GPU's and a RAC >500K I'm still slipping, so I know how you feel.
With greater participation, it's becoming more difficult to compete but it's also becoming more prestigious to be in the top 10, 50, 100, 200...
The combined RAC of the top 20 crunchers is >38M and the combined credit of the top 20 crunchers is over 11 Billion. To be anywhere near these crunchers is impressive.

In theory your GPU's should allow your RAC to be well inside the top 100 again.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

I took a break from this project but have returned. Before returning I upgraded from Ubuntu 11.10 to Ubuntu 12.10. The nVidia driver installation was difficult, a different process than on Ubuntu 11.10. I kind of think it's right but it's not working, neither here nor at PrimeGrid. I didn't change any of my task preferences on the website at first so I think it's not that.

My Event Log says:

Sun 20 Jan 2013 07:52:36 AM MST | | Starting BOINC client version 7.0.27 for x86_64-pc-linux-gnu
Sun 20 Jan 2013 07:52:36 AM MST | | log flags: file_xfer, sched_ops, task
Sun 20 Jan 2013 07:52:36 AM MST | | Libraries: libcurl/7.27.0 OpenSSL/1.0.1c zlib/1.2.7 libidn/1.25 librtmp/2.3
Sun 20 Jan 2013 07:52:36 AM MST | | Data directory: /var/lib/boinc-client
Sun 20 Jan 2013 07:52:36 AM MST | | Processor: 8 GenuineIntel Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz [Family 6 Model 42 Stepping 7]
Sun 20 Jan 2013 07:52:36 AM MST | | Processor: 8.00 MB cache
Sun 20 Jan 2013 07:52:36 AM MST | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm pcid sse4_1 sse4_2 popcnt tsc_deadline_timer aes xsave avx lahf_lm ida arat epb xsaveopt pln pts dtherm tpr_shadow vnmi flexpriority ept vpid
Sun 20 Jan 2013 07:52:36 AM MST | | OS: Linux: 3.5.0-22-generic
Sun 20 Jan 2013 07:52:36 AM MST | | Memory: 23.53 GB physical, 23.98 GB virtual
Sun 20 Jan 2013 07:52:36 AM MST | | Disk: 663.98 GB total, 626.15 GB free
Sun 20 Jan 2013 07:52:36 AM MST | | Local time is UTC -7 hours
Sun 20 Jan 2013 07:52:36 AM MST | | OpenCL: NVIDIA GPU 0: GeForce GTX 570 (driver version 304.51, device version OpenCL 1.1 CUDA, 1280MB, 1280MB available)

That last line, where it says "OpenCL:", it used to say "CUDA:" did it not? That causes me some concern but furtyher along in the same line it says "driver version 304.51" which, according an announcement from and admin I read in the News section here, is a correct driver.

And further down:

Sun 20 Jan 2013 11:14:56 AM MST | GPUGRID | Sending scheduler request: To fetch work.
Sun 20 Jan 2013 11:14:56 AM MST | GPUGRID | Requesting new tasks for NVIDIA
Sun 20 Jan 2013 11:14:58 AM MST | GPUGRID | Scheduler request completed: got 0 new tasks
Sun 20 Jan 2013 11:14:58 AM MST | GPUGRID | No tasks sent
Sun 20 Jan 2013 11:34:39 AM MST | PrimeGrid | Sending scheduler request: To fetch work.
Sun 20 Jan 2013 11:34:39 AM MST | PrimeGrid | Requesting new tasks for CPU and NVIDIA
Sun 20 Jan 2013 11:34:41 AM MST | PrimeGrid | Scheduler request completed: got 0 new tasks
Sun 20 Jan 2013 11:34:41 AM MST | PrimeGrid | No work sent
Sun 20 Jan 2013 11:34:41 AM MST | PrimeGrid | No work is available for Cullen/Woodall (Sieve)
Sun 20 Jan 2013 11:34:41 AM MST | PrimeGrid | Message from server: No work available for the applications you have selected. Please check your project preferences on the web site.

I am not seeking advice for PrimeGrid here but it's strange I cannot get any tasks from either project. I did not touch my GPUgrid task prefs at first but, I had only beta trasks selected, but now I have all types selected, short, long and beta. I've switched the "send tasks for other apps if none available for selected apps" setting on and off, clicked update on the website as well as in BOINC manager and I receive nothing. Same at PrimeGrid. I have all of their task types selected, flicked all the switches on and off in all the possible permutations yet I receive no tasks.

I am getting tasks from various CPU only projects.

In my local BOINC prefs I have selected to use GPU while computer is in use, obviously, as the client is requesting work. Under Activity I have selected Run Always and Use GPU Always.

So what's the deal? Do I now have to make a monthly cash contribution to get tasks? ;-)

Reinstall Boinc. It probably isn't recognising the driver correctly.

Sometimes it can take a while to first get a task. Maybe to do with Boinc's Scheduler request mechanism? So manually ask for new work a few times, to force the issue.

Increasing the cache a bit and restarting might also help.
304.51 is sufficient.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Reinstall Boinc. It probably isn't recognising the driver correctly.

Sometimes it can take a while to first get a task. Maybe to do with Boinc's Scheduler request mechanism? So manually ask for new work a few times, to force the issue.

Increasing the cache a bit and restarting might also help.
304.51 is sufficient.

BOINC was requesting tasks for GPU at Prime as well as here so it seems to recognize the driver. I doubt reinstalling BOINC will help. I think it's a scheduler issue. In fact I think the scheduler doesn't have the intelligence my cat's poop has. But thank the gods we have Credit New, who needs a scheduler that will fetch tasks when we have Credit New?

I increased the cache slightly and got a bunch of Prime Grid CPU tasks but no PrimeGrid GPU tasks. After I went to their site and turned off the "use CPU" setting I started getting messages saying PrimeGrid has no GPU tasks available. OK that explains part of the problem.

However, after increasing the cache and GPUgrid resource share it no longer requests tasks from GPUgrid. That is really smard! Cache is set to 4 days, I have 8 tasks in the cache, all running, it will be finished all the tasks it has in about an hour and it won't request a GPUgrid task unless I force an update. And even then it doesn't receive one. It asks for one but gets 0. Yep, my heart is filled with glee for we have Credit New and a brain dead, useless, piece of crap scheduler. Like a parachute with really cool colors and logo but no rip-cord. A 600 HP Bugatti with no wheels. Thanks David.

All my other projects are set to NNT except GPUgrid. I'm going to just let it run out of work and see what happens. I bet it won't request GPUgrid work on its own then either and if it does I bet it doesn't receive any.

I still think you should reinstall Boinc. It doesn't take long.
On Windows there is a CUDA line and then an OpenCL line in the Log Files (should be the same on Linux).
If Boinc doesn't know about the CUDA libs then it might still behave like it recognises the GPU, but won't ask/get CUDA apps and tasks.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

But it does say CUDA in the log line that starts with OpenCL, just not at the beginning of the line like it used to....

Sun 20 Jan 2013 07:52:36 AM MST | | OpenCL: NVIDIA GPU 0: GeForce GTX 570 (driver version 304.51, device version OpenCL 1.1 CUDA, 1280MB, 1280MB available)

And it requests GPU tasks from Primegrid and they are CUDA not OpenCL.

Still you could be right, maybe a BOINC reinstall will do the trick. I am down to 2 tasks being crunched which means 6 cores are unoccupied. I've detached all but one other project. There is only NumberFields left along with GPUgrid and I think this may have something to do with NumberFields@home tasks. I've noticed whenever NumberFields is in the mix the scheduler acts strange. I'm going to wait and see what happens when I finish the last NumberFields task and detach it to leave only GPUgrid in the Projects list. If it still doesn't request work on its own and still doesn't receive work when it requests then I'll reinstall BOINC.

BTW I'm not condemning NumberFields just yet, it's too early and there's not enough solid evidence.

Nope. It finished the NumberFields tasks, I detached NumberFields, still no GPUgrid tasks did I receive.

Rebooted, didn't work.

Reinstalled boinc client, it recognized the GPU as before with OpenCL at the start of the line, not CUDA, no tasks.

Rebooted and then it did NOT recognize the GPU, something strange going on.

I stopped BOINC daemon and copied the BOINC binaries into a directory in my home dir and started the client on my own account rather than the user boinc account. Still no GPUgrid tasks. Again it recognized the card with OpenCL at the start of the line rather than CUDA.

After I installed Ubuntu 12.10 and tried to install BOINC from repository, it installed all of the package(s) except libcuda1 and libcuda1-ia32. The former is the Ubuntu build of nVidia's 64 bit driver ver. 304.51, the latter is the 32 bit version. The package manager said it could not locate those 2 packages and suggested they might have been pulled out of repository. That could happen if they were found to be buggy. Or they could have been moved to a different repository for whatever crazy reason. So I googled for libcuda1, located a 3rd party package from a source I know and trust, downloaded it and installed it. Maybe it's bugs but I still think the scheduler should request work on its own when GPUgrid is the only project attached and attached just after a fresh BOINC install with no files lingering from a previous install (I deleted all those prior to reinstalling BOINC).

Anyway I've had enough banging my head on the wall. I'll do a quick check through some Ubuntu-BOINC-GPU forums to see what the scuttlebutt says and if nothing turns up then I'll roll back to Kubuntu 11.10. That worked very well before and I still have all my notes on it.

Anyone have a solution for this? Maybe an admin reset the quota? I am having this problem and would love for it to just go away.

Anyone have a solution for this? Maybe an admin reset the quota? I am having this problem and would love for it to just go away.

If you have this problem on host 327304, then you should put back the GTX980Ti, as GPUGrid is not running on GTX 1080 yet.
Your host produced too much errors, and therefore the server stopped sending work to this host.

Anyone have a solution for this? Maybe an admin reset the quota? I am having this problem and would love for it to just go away.

10 is the MAX amount of consecutive errors now. The only solution is to wait until tomorrow where you will get one task again. If you send it back "Valid" your quota will rise again so no worries if you have solved the problem of this host returning continual errors. If you return an error you will stay on 1 a day until you fix the problem.

It would be great if people wouldn't resurrect 4 year old threads, instead start a new one. It's so confusing.

The 10 max errors might prove to be a problem with these GERARD_CXCL12LOCKMONO tasks.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help