Advanced search

Message boards : Graphics cards (GPUs) : GTX 980 overheats ... after a while

Author Message
Husum / Nordfriesland
Send message
Joined: 3 Mar 16
Posts: 3
Credit: 40,627,498
RAC: 54,149
Level
Val
Scientific publications
watwat
Message 47114 - Posted: 27 Apr 2017 | 6:34:17 UTC

Good Morning from Northern Germany,

I recognized, that more or less on a regular basis - your application falls asleep. I can't give the reason - but I did the following.

I used this :=
https://www.techpowerup.com/download/techpowerup-gpu-z/

to create
http://gpuz.techpowerup.com/17/04/27/zxd.png

The important GPU Load bar stays on around 80% - then for a reason I don't know, it closes in on 100% which leads to a Card-Shutdown. Although, there is no other activity on the machine, there is no attempt for a restart - let us say - after 10 minutes. Actually, the system will sleep forever.

Please have a look.

Kind regards
René

JoergF
Avatar
Send message
Joined: 20 Apr 15
Posts: 206
Credit: 300,522,661
RAC: 1,173,140
Level
Asp
Scientific publications
watwat
Message 47115 - Posted: 27 Apr 2017 | 6:50:25 UTC - in response to Message 47114.

Good morning from Austria... :-)

would you kindly install TThrottle and record your GPU temperature so see whether or not the application overloads your card? The good thing about this tool is that you can also slow down the application in case of thermal stress and even shutdown the entire PC if the GPU temp exceeds a particular value.

Please let me know what the graph of TThrottle reads over time. Frankly I would set the temperature target of the GPU to 70°C (by Afterburner), throttle >80°C and shutdown at 90°C.

I have observed this behaviour several times with Folding@Home (possibly causing endless loops) but never had any issue with Boinc+GPUGRID ... but well, who knows.

Best wishes
Jörg

____________
I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

Erich56
Send message
Joined: 1 Jan 15
Posts: 372
Credit: 1,679,211,402
RAC: 2,917,345
Level
His
Scientific publications
watwatwat
Message 47116 - Posted: 27 Apr 2017 | 7:50:05 UTC

As I have written in another thread recently, I am observing unexplainable card shutdowns with my GTX750Ti in the Windows 10 PC since the new crunching software acemd_918.80 is in place. GPU is being overclocked very slightly only to 1150MHz.

In my case this can't have anything to do with too high temperature, since per NVIDIA Inspector I limit the GPU temp to 62°C.

The other GTX750ti which runs in the Windows XP machine, crunching software acemd_849.65, does not show this behaviour (even though it's overclocked markedly to 1250MHz).

This phenomenon of GPUs stopping crunching for unexplainable reasons has also been reported by at least one other cruncher about a week ago, shortly after the new software was distributed.

JoergF
Avatar
Send message
Joined: 20 Apr 15
Posts: 206
Credit: 300,522,661
RAC: 1,173,140
Level
Asp
Scientific publications
watwat
Message 47117 - Posted: 27 Apr 2017 | 8:00:15 UTC - in response to Message 47116.
Last modified: 27 Apr 2017 | 8:01:37 UTC

In my case this can't have anything to do with too high temperature, since per NVIDIA Inspector I limit the GPU temp to 62°C.


Well, you may be correct ... but may I add the following observation. My second gtx1070 overheats occasionally >90°C with Folding@Home although the temperature limit set to 65°C with Afterburner(!) So it seems the card can no longer check and maintain the temp limit in an endless loop.

TThrottle saved my GPU several times by PC shutdown ... and the temperature records were very revealing and informative.
____________
I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

Post to thread

Message boards : Graphics cards (GPUs) : GTX 980 overheats ... after a while