3/15/2009 1:08:31 PM|GPUGRID|Sending scheduler request: To fetch work. Requesting 49372 seconds of work, reporting 0 completed tasks
3/15/2009 1:08:36 PM|GPUGRID|Scheduler request completed: got 0 new tasks
3/15/2009 1:08:36 PM|GPUGRID|Message from server: No work sent
3/15/2009 1:08:36 PM|GPUGRID|Message from server: (reached daily quota of 4 results)
3/15/2009 1:08:36 PM|GPUGRID|Message from server: (Project has no jobs available)

Really? I'm only allowed to do 4 WUs per day? This can't be right. I detached from the project and re-attached, got the same message. Server status indicated that there's work waiting to be sent, so I don't quite get it.

Well, since March 13th you have had nothing but computation errors... That's why you have that small daily quota. If you start to send back results without error, the quota will rise again.

Most of your results show pretty different errors -

Cuda error: Kernel [shake_step_1] failed in file 'shake.cu' in line 79 : too many resources requested for launch.

ERROR: c:\cygwin\home\speechserver\gpumd2\src\pme\CPME_cufft.cu, line 50: cufftExecC2C (gridcalc2.1)

Cuda error: Kernel [pme_fill_charges_overflow] failed in file 'fillcharges.cu' in line 96 : too many resources requested for launch.

Cuda error: Kernel [frc_sum_nb_forces] failed in file 'force.cu' in line 244 : too many resources requested for launch.

Cuda error: Kernel [fft_data_swizzle_in] failed in file 'c:\cygwin\home\speechserver\gpumd2\src\pme\CPME_cufft.cu' in line 78 : too many resources requested for launch.

and so on.

Have you changed anything between March 12 and 13? New drivers maybe?

[edit] When have you rebooted your computer the last time?

---

pixelicious.at - my little photoblog

The limit 1 WU at a time per cpu core. Return (successful) results and you'll get new ones.
Edit: too late. Well, something to add: concerning your amout of computation errors it might be a good idea to check your driver and to back off your (factory?)-OC.

MrS

---

Scanning for our furry friends since Jan 2002

Believe it or not, I'm blaming SETI@home for the moment. This whole mess started when I tried their "optimised" Astropulse files. Numerous blue screens, random system restarts, etc. I blew away everything related to BOINC, but the problems are continuing on a lesser scale. Their current Astropulse seems to be REALLY problematic and I shut it off this morning as a result. The unplanned reboots due to BSOD were likely to blame for the computation errors. I had thought the ones that completed last night were good, but I guess not. System runs without issue with BOINC not processing anything, so it's definitely something in there.

In any case, I have to wait until tomorrow to get new WUs to try?

In that case I'd suspend SETI and see how GPU-Grid does without being disturbed. There have been other reports of computation errors after SETI ran. And yes, you'll probably have to wait 24h for the next try.

MrS

---

Scanning for our furry friends since Jan 2002

Have you checked your NVIDIA Driver version lately? I run SETI alongside GPUGrid, and noted that SETI tends to be more sensitive to latest NVIDIA driver than GPUGrid with the optimised third party apps (as opposed to the standard Project stock apps). Once my drivers were updated I had no issues with either project.

Regards
Zy

I did a clean install of the latest WHQL nVidia drivers yesterday afternoon. After my experience with the optimised client for SETI, it's very unlikely I'll give it another shot. I haven't gotten a CUDA WU from SETI in a couple weeks due to scheduler issues on their end, so I don't think it's a video issue. I'm just hoping this issue is behind me, looks like I'll find out tomorrow.

Keep an eye on the temps of cpu and gpu also.
See if you have enough diskspace just to make sure these are not being the problems

Certain SaH task deaths seem to leave the GPU in a bad state. RUnning SaH alone means that all the rest of the SaH tasks are run into Computation error till the system is restarted. In some cases it has to be a "cold" start with the system turned all the way off.

As others have suggested, try running without SaH for a bit and see if that helps ...

Yea, it was the overclock. Looks like I need to update my preferred stability testing protocol in the thread I responded to last week. My previous method missed some miscalculating by the CPU at the speeds I was pushing. SETI@home, GPUgrid, and the other projects are all processing without error now. Interesting that some projects were more tolerant of the miscalculations than others.

Yea, it was the overclock. Looks like I need to update my preferred stability testing protocol in the thread I responded to last week. My previous method missed some miscalculating by the CPU at the speeds I was pushing. SETI@home, GPUgrid, and the other projects are all processing without error now. Interesting that some projects were more tolerant of the miscalculations than others.

Makes sense that it was a O/C problem, specially since you previously pointed out, that the troubles started when you began using s@h optimized apps.

Dunno how that might have affected GPUgrid, but s@h optimized apps. stress the comm. bus {FSB} quite a lot more (due to heavy instruction utilization), perhaps the reason behind your BSOD's and failed tasks of several projects, specially since they run on your 4core X9650 simultainously ~~

Anyway a slight back-off on your high O/C apparently solve the issue, so we're happy for ya^^

Who said anything about backing off? Up the voltage, baby! :)

1.312V under load and it's happy as a clam at 3.80 GHz. I don't think 4.0 is going to be possible with this motherboard/CPU pushing sane voltages, which is a bit of a shame.

Oh well, yes, more juice ^^
Anyway are you worried about the lil 45nm silicon to die on ya at higher v's? :)
Don't be, I'm pushing 1,45v's to my E8200 for over 14months now 24/7 no issues whatsoever and I don't expect any, as long as temp's are kept in check.

hf mate~