GTX 980 FTW Edition idles at 25% power usage, 60C

Apr 8, 2018
7
0
10
Cross posting this from a reddit post that's just getting zero attention.

Computer Type: Desktop, custom build.

GPU: GTX 980, 4GB VRAM, no overclock

CPU: Interl i7 6700k 4ghz

Motherboard: Gigabyte ATX DDR4 Intel LGA 1151 SATA E (6Gbit/s) Motherboards (GA-Z170X-Gaming 7.) Can't seem to find the BIOS version, but it was whatever came on there, never updated/flashed/whatever.

RAM: 16GB, I think corsair.

PSU: Cooler Master Silent Pro M - 1000W, 12v rail, not sure about the output, kinda confused by all this.

Operating System & Version: Windows 10 build 16299, upgraded from 7.

GPU Drivers: 388.43

Description of Problem: Provide as much info about the issue as you possibly can, images and videos can be provided as well.

Troubleshooting: About a week ago I noticed my 980's fan was kicking in while idling, which I thought was strange, since it doesn't normally do that unless my room temp is hot. It appears my card was drawing 25% and heating up to around 60C. I went through a lengthy process of trying to figure this, which included:

Removing Precision and Afterburner
Completely reinstalling GPU drivers (from the latest version.) -Turning hardware acceleration off in anything that uses it
Examining Process Viewer and task manager to see if anything stood out as strange (nothing did)
Running a malware scan (nothing found)
Running CCleaner.

If I start my computer up from a full shutdown it will draw about 8% power while idling for some time, usually around an hour or so, then at some point it just jumps up to 25% usage and holds there, even if I shut every program I can find down.

The programs I would usually run at any given point in time are Firefox or Opera, Discord, and Steam. Again, with all of those off, the GPU in GPU-Z will show a 25% draw and 60C idle temps which seems abnormal to me. I'm not sure where else to look or what could be going on and I worry about this possibly damaging the card in the long run.

The only major change I can think that I made recently was installing a Huion drawing tablet, but it's not plugged in and I'm not actively running any of its software as far as I can tell.

http://gpuz.techpowerup.com/18/04/07/n4v.png

The last thing I tried was downgrading my driver to 388, which I believe is the factory default. For a while this worked, I was idling at 8-11% usage which was fine, with temps between 30-40C. The higher end of that is likely due to ambient temperature in my home. However, after returning to my room an hour later, the CPU was hanging at 25% again, 60C, with nothing running. I used Invidia Inspector and set up the "Multi Display Power Saver" option as instructed in a thread elsewhere for a similar issue. So far that seems to be keeping it down, and I read it should still draw more power when it's needed, but I don't want to say the issue is totally resolved just yet.

I'm wondering what might cause this and if there's something I should be doing to prevent it in the future. Is it normal for this card to behave this way, is using the "Multi Display Power Saver" option a good fix, or should I be looking into something else?
 
Apr 8, 2018
7
0
10
Trying simple color doesn't seem to make a difference.

honestly I can't think of much I have changed. I got a Huion drawing tablet, so that acts as a second monitor at times, but don't always leave it plugged in. I had updated to new drivers about a week before that, so I had the latest Nvidia was offering.

The only other major change I can think of is that my computer crashed while playing Injustice 2 recently. There's some bug that I had to verify the cache to solve, but the first time it just took my system with it. I had to verify the disk and all that afterwards.

I have had this happen a few times in the past but the solution was always more simple. Switch to adaptive power, or reinstall the drivers. It's never done this with such consistency.

I had MSI but removed it years ago and mostly used Precision to check GPU temps and usage, I've since switched to GPU-Z.

Side note, I put my computer to sleep and woke up a while ago, started it up again and threw on a video and got another hour of rest. When I woke up, the memory clock was at about half usage. I closed everything out and it wouldn't go down. I noticed Nvidia Inspector wasn't in my task tray so I opened it again and after doing so memory usage went back down to idle levels. The power draw during all of this though was 10% which seems fine, and the card hadn't heated up beyond 45C (again, a bit warm in my room, so that's no doubt part of it.)
 
Apr 8, 2018
7
0
10
Blew out dust yesterday. Not sure how to change thermal paste on the GPU, or thermal pads, though. Would that cause it to begin drawing more power, even though the fans aren't really running? Main thing is it seems like it's in use. Memory goes up, draws more power, then the card heats.

Should still do what you said though, I live in the desert and the Summers get hot, it would certainly help.
 
Just before you do this, have you done clean GPU driver installation via DDU (Display driver uninstaller)?

Download it and latest drivers (Select drivers and windows dont do automatic).
Now I suggest you Shift+Restart method. https://www.digitalcitizen.life/4-ways-boot-safe-mode-windows-10

After doing "Safe mode" run DDU as admin, then Select Nvidia and Clean+restart and it should restart automaticly, now install your Nvidia drivers and see if there is any change.


Well, as i said there is no need to change thermal paste if the GPU isnt older than 3 years (if bought from shop when it came out) because the thermal paste dries over the years, but for thermal pads they are replaceable (usually every 5 years it should be changed)

Well if you have time and will i suggest you to do it.

Any weird virus or something? Download Adwcleaner and run it and same goes for malwarebytes (just to be sure).

One weird thing, you are not using latest windows? (maybe do an update?)

Any perfomance drop in games?
 
Apr 8, 2018
7
0
10
Just ran through the steps you outlined, using DDU and installing the new drivers. Should I also switch to adaptive rather than optimal in the Nvidia control panel?

Too soon to tell if that did the trick though. Like I said, this issue usually pops up after idling for a while. Strangely enough it seems if it doesn't just do it on start up, it does it mostly after being away and not doing anything, or after waking from sleep.

I've run Malwarebytes and Windows Defender over two different sessions, both came up clean. Just ran adwcleaner and it also says nothing was detected.

Checking for updates now. Right now it says it's downloading an nvidia update which seems odd to me.

There hasn't really been any performance drop in games, but I haven't played around in much. Any lag I have experienced is to be expected, though.

Edit: Says my computer is up to date as far as Windows goes.
 
Apr 8, 2018
7
0
10
So basically all day it was fine. I just accidentally closed GPU-Z (hate how if you minimize it doesn't just drop to the task bar) and when I pulled it back up everything jumped to 23% usage and now it's hanging there again. I ran Nvidia Inspector again and that dropped everything back down. I really don't understand what the problem is... I feel like I've done about everything I can, and I'm really wondering at this point if this is just how the card operates, or - in the worst case scenario - if the card or PSU are faulty.

Edit:

Decided to turn multi-display power saving off from the Inspector control panel. This immediately causes the memory usage and power draw to spike. However, if I turn it on, then kill the process, the GPU-Z readings hold at normal idle levels.

 
Weird, seems like driver issue or software, there must be conflict between them.

"I've posted this a few times so hopefully it can help you out.

Nvidia inspector enables your system to downclock automaticaly on idle through multi display power saver function. Open nvidia inspector and right click "show overclocking" and select multi display power saver. This program allows you to force p8 state unless a program in the exceptions list is running (you will need to add these one by one).

From what I've read, the nvidia cards used to downclock just fine on idle, but some users were experiencing extreme screen flickering on their high refresh rate monitors. Nvidia decided to just force a p0 state on idle as a blanket solution to these issues."

Sorry but i've Never used Nvidia inspector,i was doing a fix 10 years ago on my old system with Hitman.
 
Apr 8, 2018
7
0
10
Hm, I haven't read about that option in Inspector. From what I read, the dual monitor power save option still allow the card to clock up when it needs to, but fixes the issue of not downclocking while idle. I tried a game and the card does seem to start clocking up as normal when it's put to use, and does downclock as it should, provided Inspector's Dual Monitor Power Save is enabled. The odd thing to me is killing the process in task manager almost makes it so it's still active, when it's not.

My first assumption was that this had to be a driver issue. I can't fathom what software may be conflicting though. Is there a method that could narrow it down?

In the meantime, the monitor power save SEEMS to be working, and if it's not going to cause any adverse issues or weirdness, I'm comfortable letting it do its thing.