Piles of error evolve suddenly in a certain day

Questions and Answers : GPU applications : Piles of error evolve suddenly in a certain day
Message board moderation

To post messages, you must log in.

AuthorMessage
Tse Kin Fai
Volunteer tester

Send message
Joined: 13 Jan 10
Posts: 9
Credit: 130,436
RAC: 0
Hong Kong
Message 967333 - Posted: 31 Jan 2010, 18:22:20 UTC
Last modified: 31 Jan 2010, 18:58:09 UTC

My computer:
http://setiathome.berkeley.edu/show_host_detail.php?hostid=5262954

There isn't this kind of error before,
I figured out what the sign of the error is, would someone expert in this help me to solve the problem?

The situation when the CUDA is functioning normally is:
The GPU usage sensor should record 0% usage(due to the problem of NVIDIA driver) and the CUDA application usually occupy 1 of the CPU core but not up to 100% of that core

The sign for a CUDA WU to corrupt is:
GPU usage rise to over 80%
occupy more system memory suddenly
Forced a core up to 100%
The screen suddenly flash(blacked out for about 0.1s)
These cause my computer to have a really low framerate and lagging windows.

What is running on my computer is just the antivirus, a firefox browser, thats all and can never cause a sudden rise in GPU usage

Here are the links to those corrputed job, all ended with unspecified launch failure.

http://setiathome.berkeley.edu/result.php?resultid=1500278988
http://setiathome.berkeley.edu/result.php?resultid=1500687713
http://setiathome.berkeley.edu/result.php?resultid=1500950753
http://setiathome.berkeley.edu/result.php?resultid=1500934055
http://setiathome.berkeley.edu/result.php?resultid=1500933900
http://setiathome.berkeley.edu/result.php?resultid=1500932161
http://setiathome.berkeley.edu/result.php?resultid=1500931013
http://setiathome.berkeley.edu/result.php?resultid=1500902998
http://setiathome.berkeley.edu/result.php?resultid=1500900453
http://setiathome.berkeley.edu/result.php?resultid=1500900451
http://setiathome.berkeley.edu/result.php?resultid=1500900449
http://setiathome.berkeley.edu/result.php?resultid=1500889751
ID: 967333 · Report as offensive
Profile Fred J. Verster
Volunteer tester
Avatar

Send message
Joined: 21 Apr 04
Posts: 3252
Credit: 31,903,643
RAC: 0
Netherlands
Message 967366 - Posted: 31 Jan 2010, 22:32:50 UTC - in response to Message 967333.  

Hi, your card isn't OC'ed, those errors are strange, at least.
Card not being or getting too hot?
Install the latest drivers. And try to 'test it' , Si Soft Sandra, Everest, etc.


ID: 967366 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 967369 - Posted: 31 Jan 2010, 22:49:59 UTC - in response to Message 967366.  

Frted, he's got the latest driver, 196.21. It might be better if he tries rolling back to the 191.xx driver. The 196.21 driver has been causing problems for some people.

Fai, when did you get the newest driver? I see your problem has only been happening the last couple of days. You may want to install the newer 2.3 DLLs too. You can get them here http://lunatics.kwsn.net/index.php?module=Downloads;sa=dlview;id=208 You may have to join the lunatic's site to get them.


PROUD MEMBER OF Team Starfire World BOINC
ID: 967369 · Report as offensive
Tse Kin Fai
Volunteer tester

Send message
Joined: 13 Jan 10
Posts: 9
Credit: 130,436
RAC: 0
Hong Kong
Message 967421 - Posted: 1 Feb 2010, 8:53:01 UTC

I've got the driver for about two weeks and it doesnt cause any problem until that time.
And the display card should not be over heated, as some time before, the temperature is even higher than that i have now, it is still below 60 degree Celcius, but it is somehow strange that CPU is 5x degree Celcius
ID: 967421 · Report as offensive
Tse Kin Fai
Volunteer tester

Send message
Joined: 13 Jan 10
Posts: 9
Credit: 130,436
RAC: 0
Hong Kong
Message 967452 - Posted: 1 Feb 2010, 14:51:05 UTC - in response to Message 967421.  
Last modified: 1 Feb 2010, 15:26:12 UTC

finally i think it is the time for the card to go back go the factory

The same thing happens to my previous display card, the problem is there is some memory locations corrupted causing all those errors. Downgrading the driver to 191.07 doesn't help...

Thank you all for suggesting the possible situations but I will be off from CUDA until the card is fixed. I now have to remove the display driver and use the CPU to do the graphics work.
ID: 967452 · Report as offensive
Profile Fred J. Verster
Volunteer tester
Avatar

Send message
Joined: 21 Apr 04
Posts: 3252
Credit: 31,903,643
RAC: 0
Netherlands
Message 967471 - Posted: 1 Feb 2010, 16:00:27 UTC - in response to Message 967452.  

Hi, could be a VRAM error (defect memory chip), causing this.
Try to test it yourself or send the card back to factory/seller.
I assume, it's still erroring?


ID: 967471 · Report as offensive

Questions and Answers : GPU applications : Piles of error evolve suddenly in a certain day


 
©2026 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.