Graphics card idling at 25 degrees above ambient

Tanyac

Reputable
For the last 24 hours I have been unable to get my GPU temp below 37 degrees Celsius (19 degrees above ambient), running the fan at 70%. When idling with zero activity on the card or my system, with the fan at 0% (It doesn't automatically switch on), the GPU sits at 48 degrees Celcius (30 degrees above ambient).

I've cleaned out the system (The case temp is 25 degrees, and CPU is 31 degrees - Both of which are higher than normal - perhaps they are suffering the effects of the excessive GPU temps?)

I'm hesitant to RMA the card unless I absolutely have to as the warranty process will take 6 - 12 weeks here. I don't have a spare card and I depend on the PC for all my work, so I can't be without it for more than 24 - 48 hours.

The Card is a ASUS GTX 980 Ti, about 14 months old.

Any suggestions?
 
Solution
How recent are your drivers? And what is the GPU clock rate when no other applications are running?

I ask because there has been an issue recently where the GPU does not down-clock after playing certain games (Doom in my case). Mine would sit @1417Mhz with no apps running. Sometimes the condition seemed to persist across reboots which was really weird.

I uninstalled drivers and installed a version from June and that fixed it for me.
How recent are your drivers? And what is the GPU clock rate when no other applications are running?

I ask because there has been an issue recently where the GPU does not down-clock after playing certain games (Doom in my case). Mine would sit @1417Mhz with no apps running. Sometimes the condition seemed to persist across reboots which was really weird.

I uninstalled drivers and installed a version from June and that fixed it for me.
 
Solution


Perhaps not. But 2 days ago the temps were sitting at 6 - 8 degree above ambient with the ambient at 18 degrees. Over the last 48 hours this has increased to 30 degrees over ambient. What has changed - Nothing!

I haven't changed a thing on my system. It's kept clean inside all of the time, yet it jumped 22-24 degrees. Surely that's a clear sign that something has, if not gone wrong, then aat least "changed" with the card itself?
 


According to GPU Tweak..

GPU Boost Clock = 1291 MHz
GPU Voltage = 1163 mV
Memory Clock = 7200 MHz

Was there another value you wanted?

GPU Tweak is set to gaming mode, which is the default, and I haven't changed it.

Graphics driver version is 384.76, which it's been on since 11th of July (Released on June 29th)

I notice there is now a 385.69. I will try a few different versions and see what happens.


 


If that's the current clock rate (GPU and memory) then yeah, it's not really idling. I don't recall the exact driver version I used but it was from sometime in June. Maybe the one right before yours?
 


Driver version 382.53 from 9th June 2017 resolved the issue. I tried several driver versions, and as you said, the one from June was likely the solution.

Why this took two months to surface is bizarre, but at least I know every version since 382.53 has a problem with my system.

thanks for your assistance.

ps. 1291MHz is the default clock speed out of the box.