These have just started to appear on my 2x260gtx,

23 Jan 2011 11:23:04 UTC Error while computing 839.27 7.00 0.05 --- MilkyWay@Home v0.50 (cuda_opencl)

The cards are well within running capabilities, well cooled, and have been working fine since dec 10.

anything to do with the back log of wu waiting for validation?

thanks

Hi,

If it is only one or two who error-out then I would not bother. Very occasionally I have a WU that errors, on several projects. If all error-out than you could reboot the system or update the drivers.

I had a problem with my GTX286 (blue screen and reboot while running MW).
I have updated to the latest driver and now the WU's run complete.
Just an idea, but I don't know if this works for you.

---

Greetings from,
TJ

Hi

No there is a list of them, going back to yesterday, and current wu are also showing error, there is no error report in boinc manager, I am tracking the time on the site and then in boinc.

Latest error

23 Jan 2011 12:33:15 UTC Error while computing 832.21 6.93 0.05 --- MilkyWay@Home v0.50 (cuda_opencl)

from boinc


23/01/2011 12:28:13 World Community Grid Finished upload of X0000060460364200511021055_0_0
23/01/2011 12:30:14 Milkyway@home Sending scheduler request: To fetch work.
23/01/2011 12:30:14 Milkyway@home Requesting new tasks for GPU
23/01/2011 12:30:16 Milkyway@home Scheduler request completed: got 0 new tasks
23/01/2011 12:30:16 Milkyway@home Message from server: No work sent
23/01/2011 12:30:16 Milkyway@home Message from server: (reached limit of 36 tasks in progress)
23/01/2011 12:33:06 World Community Grid Computation for task X0000060460148200511021058_1 finished
23/01/2011 12:33:06 World Community Grid Starting X0000060450346200511031437_0
23/01/2011 12:33:06 World Community Grid Starting task X0000060450346200511031437_0 using hcc1 version 608
23/01/2011 12:33:08 World Community Grid Started upload of X0000060460148200511021058_1_0
23/01/2011 12:33:12 World Community Grid Finished upload of X0000060460148200511021058_1_0
23/01/2011 12:35:06 Milkyway@home Sending scheduler request: To fetch work.
23/01/2011 12:35:06 Milkyway@home Requesting new tasks for GPU
23/01/2011 12:35:08 Milkyway@home Scheduler request completed: got 0 new tasks
23/01/2011 12:35:08 Milkyway@home Message from server: No work sent
23/01/2011 12:35:08 Milkyway@home Message from server: (reached limit of 36 tasks in progress)
23/01/2011 12:35:39 Milkyway@home Computation for task de_separation_19_3s_fix_2_257996_1295769003_1 finished
23/01/2011 12:35:39 Milkyway@home Starting de_separation_19_3s_fix_2_280334_1295771644_0
23/01/2011 12:35:39 Milkyway@home Starting task de_separation_19_3s_fix_2_280334_1295771644_0 using milkyway version 50
23/01/2011 12:35:48 Milkyway@home Computation for task de_separation_23_3s_fix_1_274603_1295771099_0 finished
23/01/2011 12:35:48 Milkyway@home Starting de_separation_19_3s_fix_2_281584_1295771817_0
23/01/2011 12:35:48 Milkyway@home Starting task de_separation_19_3s_fix_2_281584_1295771817_0 using milkyway version 50
23/01/2011 12:36:48 Milkyway@home Sending scheduler request: To fetch work.
23/01/2011 12:36:48 Milkyway@home Reporting 2 completed tasks, requesting new tasks for GPU
23/01/2011 12:36:49 Milkyway@home Scheduler request completed: got 0 new tasks
23/01/2011 12:36:49 Milkyway@home Message from server: No work sent
23/01/2011 12:37:54 Milkyway@home Sending scheduler request: To fetch work.
23/01/2011 12:37:54 Milkyway@home Requesting new tasks for GPU
23/01/2011 12:37:56 Milkyway@home Scheduler request completed: got 0 new tasks






You can see at 12.33 the task finished fine with no errors in boinc, so why is it showing as error on milkyway site?


cheers

Is sent time on stats page from server to computer?

The tasks in your error list seems to be all from the series de_separation_23_3s_fix_1. All of them with errors like
Chunk estimate: 21
Num chunks: 20
Added area: 0
Effective area: 280000
Global dimensions not divisible by local
Failed to find good run sizes
Failed to calculate integral 1
So I guess it's related to the formula problem Matt mentioned.

Since the app isn't erroring out but finishing with a result (here 'failed'), your boinc client doesn't see an error; the validator on the server is checking your result, sees that it failed to calculate the last chunk of data und marks it as computing error. In real it is a problem of the app with some WU types only shortly discovered.

Many thanks for that info Len LE/GE


The 2x260gtx has been previously running fine, the only thing that is different was I forgot to change the power management in the nvidia control panel after reformat, just changed to prefer maximum performance. I have stopped on these 260gtx for now also on milkyway, my other ati cards look to be running ok


cheers

I have to report similar problems with the 23 series similar to the problem below

http://milkyway.cs.rpi.edu/milkyway/result.php?resultid=297816174

the 23 series is the only type failing.
Q9300
gt260
boinc 6.10.58
app ver MilkyWay@Home v0.50 (cuda_opencl)

I have aborted all 23 series
addendum : all 23 series fail !!!!

---

OWN every thing I need
EARN.. enough to live !!!
WANT a solar array on the roof so I can run a BOINC farm( DREAM on!!)
NO wife
NO kids
NO troubles

all 23 series(open cl) workunits downloaded are also all failing with computing error message. have aborted all downloaded workunits and suspended Milkyway.

Admin appears to be aware of this problem see this post

http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=2107&nowrap=true#45764