[SOLVED] 91°C GPU temperature with 2% GPU usage

[alphardxyz]

https://imgur.com/a/bu9g8Cc

View: https://imgur.com/a/bu9g8Cc


This is my GPU while having Doom eternal open in the main menu (not even playing). As you can see from the image, the temperature is 91°C while the GPU usage is consistently below 2%. At this point the GPU starts to throttle so any slightly demanding game is unplayable.

How is this possible? The GPU is barely doing anything, yet it's frying . By the way, this temperature is reached less than 2 minutes after booting up the game.

I've already tried replacing the thermal paste correctly, cleaned up the fans, updated GPU drivers, etc. Also tried undervolting the GPU. No measurable change whatsoever. Seems like the problem is more serious?

For reference I'm on a Dell G5 15 , processor is an intel i7 8750H, 16 GB Ram, GTX 1060 Max Q.

Is it time to throw it in the garbage bin?

 

[Adam65]

Hey, when did you refresh your thermal paste? If you're not sure, check this one out ;


https://www.amazon.com/Thermal-Grizzly-Kryonaut-Grease-Paste/dp/B00ZJSF5LM

It's one of the best ones. I think the top1 one, 12 W-m/K, with this installed on your GPU, it wont even reach 60 degrees,

If your thermal paste is new. Then there must be a software issue

 

[alphardxyz]

Hey, when did you refresh your thermal paste? If you're not sure, check this one out ;


https://www.amazon.com/Thermal-Grizzly-Kryonaut-Grease-Paste/dp/B00ZJSF5LM

It's one of the best ones. I think the top1 one, 12 W-m/K, with this installed on your GPU, it wont even reach 60 degrees,

If your thermal paste is new. Then there must be a software issue

I replaced it yesterday, it was a good brand, and it didn't make any difference, so that's not the issue.

I'm not sure what you mean by 'software issue'. How does a software make my GPU overheat?

 

[Adam65]

Well, my GTX 760 X2 (rog mars) was sitting at 60 degrees or so at idle, but was doing nothing, i tried using DDU, and installing an old model of nvidia drivers, it was so cold ! (20-30 degrees), and then i tried to download latest driver again, problem was fixed

 

[Adam65]

By the way, I'm scared when my GPU hits 80 degrees(im poor and have no money to but another one if it breaks), so you might want to close that laptop till the problem gets solved, try using DDU and a fresh driver

 

[Phaaze88]

For it to be that high typically means poor paste application or loose/improper heatsink mount.
Even in a stuffy laptop, it shouldn't be that high just idling on the desktop. You did something wrong during the laptop service.

 

[digitalgriffin]

https://imgur.com/a/bu9g8Cc

View: https://imgur.com/a/bu9g8Cc


This is my GPU while having Doom eternal open in the main menu (not even playing). As you can see from the image, the temperature is 91°C while the GPU usage is consistently below 2%. At this point the GPU starts to throttle so any slightly demanding game is unplayable.

How is this possible? The GPU is barely doing anything, yet it's frying . By the way, this temperature is reached less than 2 minutes after booting up the game.

I've already tried replacing the thermal paste correctly, cleaned up the fans, updated GPU drivers, etc. Also tried undervolting the GPU. No measurable change whatsoever. Seems like the problem is more serious?

For reference I'm on a Dell G5 15 , processor is an intel i7 8750H, 16 GB Ram, GTX 1060 Max Q.

Is it time to throw it in the garbage bin?

Restore voltages to stock. Also I don't think you'll hit 3GHz. lol. You might hit 1.8 IF you are lucky.

 

[Phaaze88]

Restore voltages to stock. Also I don't think you'll hit 3GHz. lol. You might hit 1.8 IF you are lucky.

I don't see where the OP touched the core voltage?
That 3ghz is the memory clock, isn't it?

 

[digitalgriffin]

I don't see where the OP touched the core voltage?
That 3ghz is the memory clock, isn't it?

I might have misread that. The screen cap was small.

 

[Adam65]

Yeah, he replaced the paste yesterday, this means he put not too much paste, or one of the screws are loose, software stuff cant matter that much, i mean 91 degrees at idle,

 

[Phaaze88]

Nah, if it was a software issue, it would be throwing out some crazy numbers, like 200C, or -80C; values that basically wouldn't make any sense.
This looks like poor cooler contact.
Example: after applying paste, the OP secured the heatsink by fully twisting one screw at a time, instead of twisting a little bit and moving to the next screw, repeating until all screws stop.
Or the paste didn't make a complete spread over the die under the mounting pressure.

 

[Adam65]

I usually screw the bottom left one a little tiny bit, enough to hold them together, then the top right one, and then screw the others in place, then screw the little loose screws (Don't screw them too hard)

 

[alphardxyz]

Nah, if it was a software issue, it would be throwing out some crazy numbers, like 200C, or -80C; values that basically wouldn't make any sense.
This looks like poor cooler contact.
Example: after applying paste, the OP secured the heatsink by fully twisting one screw at a time, instead of twisting a little bit and moving to the next screw, repeating until all screws stop.
Or the paste didn't make a complete spread over the die under the mounting pressure.

Yeah that's what I did, I think. I just followed my laptop's handbook, and it only said to twist them in the correct order.

The thing is, I had this problem before replacing the thermal paste. So the heat sink must have been already badly secured, before I touched anything.

My CPU kind of overheats too, now that I think about it, though not as bad as the GPU, but it regularly reaches 90° as well. Is this suggestive of a problem with the heat sink?

 

[alphardxyz]

For it to be that high typically means poor paste application or loose/improper heatsink mount.
Even in a stuffy laptop, it shouldn't be that high just idling on the desktop. You did something wrong during the laptop service.

It wasn't idle, I was running a game (Doom eternal). I think the 2% gpu usage must be some kind of bug with the task manager, because it doesn't make sense to only have 2% while running that game. Still, it reaches this temperature after just 2 minutes of starting the game.

 

[Phaaze88]

Check usage with Afterburner, not Task Manager.

My CPU kind of overheats too, now that I think about it, though not as bad as the GPU, but it regularly reaches 90° as well. Is this suggestive of a problem with the heat sink?

Negative. This is typical of cheaper gaming laptops, unfortunately.
Cooling is one of, if not the first place, where costs are cut with these devices.

 

[alphardxyz]

Check usage with Afterburner, not Task Manager.


Negative. This is typical of cheaper gaming laptops, unfortunately.
Cooling is one of, if not the first place, where costs are cut with these devices.

I get it, but it wasn't like this when I first bought it. It worked fine for the first year or so, then it randomly started with this crazy gpu overheat problem. That's why I think some hardware must have deteriorated.

 

[Adam65]

Was this your first thermal paste changing, i mean did the laptop evet got opened

 

[Phaaze88]

Old thermal paste dried up?

Torn thermal pads during self service?

Another possibility - and I'm not sure if this is actually a thing, or marketing fluff - heatpipe dry-out.
https://www.icegiantcooling.com/blog

 

[spacejunk]

Sounds like a heatsink mounting issue. I'm not sure about laptops as I've never owned one, but the easiest way to diagnose/eliminate the heatsink is to apply mild pressure to the heat-sink, while monitoring temps.

I had this problem with my gtx 980. I eventually resorted to using cable ties to clamp the HSF to the PCB, that's because I am too lazy to take it apart. BTW I never touched the 980, it was essentially a new card.

 

[alphardxyz]

Old thermal paste dried up?

Torn thermal pads during self service?

Another possibility - and I'm not sure if this is actually a thing, or marketing fluff - heatpipe dry-out.
https://www.icegiantcooling.com/blog

This is how it was before i replaced the thermal paste. Do you think those thermal pads could cause problems?

https://imgur.com/a/iReBJnD

View: https://imgur.com/a/iReBJnD

 

[alphardxyz]

Sounds like a heatsink mounting issue. I'm not sure about laptops as I've never owned one, but the easiest way to diagnose/eliminate the heatsink is to apply mild pressure to the heat-sink, while monitoring temps.

I had this problem with my gtx 980. I eventually resorted to using cable ties to clamp the HSF to the PCB, that's because I am too lazy to take it apart. BTW I never touched the 980, it was essentially a new card.

Thanks, but I'm not sure how to apply pressure to the heat sink while the laptop is running. Maybe I'll just try to replace the heat sink again, twisting the screws little by little.

 

[Phaaze88]

The thermal pads don't become a problem unless you tore them during the tear down of the laptop.

 

[Adam65]

Which thermal paste did you apply,
And next time please use a small amount, what i saw on photo was brutal amount of paste