GPU temps reach hot temps (87 degrees C)

[Iggy Wolf]

So I just recently bought a new GPU (Gigabyte RTX 2060 Super OC). My previous one was an ASUS STRIX GTX 970 OC. I have a Corsair 750M in the case and two case fans for cooling (one is 120 mm). I was just playing Rise of the Tomb Raider and despite the fact that with my old card, the computer case didn't get hot or heat up, for some reason, after installing this card, after a couple of hours the computer case started to get hot and I noticed it was particularly hot where the GPU is.

The CPU temps were only showing 55-60 C. The GPU monitor showed it got as high as 87. Why would the new card get hotter than the old one despite me not changing anything else about the system? The card only requires 550 W so my 750 W PSU is more than enough, and the air flow is usually good. The room itself is hot though due to outside temperatures being hot lately but I still don't see why that would heat up the card. Also, I didn't install the Aorus Engine that comes with the card because I didn't care for controlling fan speed or overclocking. Is the program necessary for the GPU fans to properly adjust to high temps?

[Iggy Wolf]

All I'll say is that the heat is only problematic to me and not the GPU because the rest of the case ends up warming up as a result, increasing the CPU's temperature as well as making my WiFi USB dongle warm up. It's an aluminum case, so the whole thing ends up heating up and feels like a BBQ grill.

Granted, right now, this heat wave of 100+ temperatures outside, and the AC inside barely keeping the room cool contributes to the higher ambient temperature, but like I said, I expect the card to increase in temps on full load, not simply upon starting a game. I'll test all my games to see whether it's dependent on specific games or is simply an issue with the card's cooling.

Edit: After testing another game, the thing I noticed instantly is how quickly the GPU begins to hit the temp limit (83 C). The base clock is 1500 but the boost clock is 1750 and yet I noticed during the game it alternates between 1750, 1780, 1730, as if the card is throttling itself every time it hits 83 and goes back to 81 only to increase back to 83 again. The fan's even working at 90%. Quite frankly, I'm not sure what kind of fan curve I'd have to create to prevent it from exceeding 76. It doesn't seem like the fan is able to cool it fast enough, and the only reason the temp doesn't go higher is because it throttles itself back. That's inefficient cooling no matter what case I have, since the card immediately gets hot very quickly.

[[☥] - CJ -]

Thats the boost clock, its normal.
Temps dictate how high or low the boost clock will be
You could just use Afterburner to raise the temp limit, unless its already maxed.

Did you uninstall the previous GPUs drivers and install the latest?

i still think what i suggested should be attempted, but i also think the difference in coolers could be the cause.
-

The other possibility is...
The factory did a sucky job on applying the thermal paste
when i redid the paste on my 1060, it dropped load temps up to 30C, yes, UP TO 30C
And now because of such a large difference in temps i was able to do a pretty large overclock and it still doesnt go higher than 60-62C max load.

So thats something to consider as well

[Iggy Wolf]

Publicado originalmente por ☥ - CJ -:

Thats the boost clock, its normal.
Temps dictate how high or low the boost clock will be
You could just use Afterburner to raise the temp limit, unless its already maxed.

Did you uninstall the previous GPUs drivers and install the latest?

i still think what i suggested should be attempted, but i also think the difference in coolers could be the cause.
-

The other possibility is...
The factory did a sucky job on applying the thermal paste
when i redid the paste on my 1060, it dropped load temps up to 30C, yes, UP TO 30C
And now because of such a large difference in temps i was able to do a pretty large overclock and it still doesnt go higher than 60-62C max load.

So thats something to consider as well

Thanks. As I've never really taken off the fans off my cards, how can I tell then whether the thermal paste is applied or not? Also, since I bought the card within the last 30 days, could I just simply RMA it for a new one, which hopefully will have a better job done on the thermal paste? I've never applied thermal paste myself, much less taken apart a card. It's still under warranty and I feel like reapplying the past would void the warranty.

[[☥] - CJ -]

Its not that the paste wasnt applied, its just the factory did a poor job applying it.

It wouldnt void the warranty, repasting is allowed.

But yeah, just contact Gigabyte and apply for an RMA, or ask them for support at least.
Before all that, i would give what i suggested in my 1st post a try first

[Iggy Wolf]

Publicado originalmente por ☥ - CJ -:

Its not that the paste wasnt applied, its just the factory did a poor job applying it.

It wouldnt void the warranty, repasting is allowed.

But yeah, just contact Gigabyte and apply for an RMA, or ask them for support at least.
Before all that, i would give what i suggested in my 1st post a try first

Thanks. I was actually wondering, how would I go about uninstalling the old drivers and reinstalling with completely clean drivers? When I was installing the drivers for this card, I didn't click the "Clean install" option, so I think the old drivers are still present in some manner on the PC. I don't know if you or someone else mentioned to use DDU, restart the PC in safe mode, and then reinstall the drivers, or use DDU while in safe mode and then restart regularly and install the new drivers.

[Iggy Wolf]

Ok, so what does it mean when your card is practically EXCEEDING the max temp limit? MSI Afterburner shows me that its temp limit is 83 C. It reached 86 C in-game, the fans started to run at 100% (the graph showing them attempting to run at 112%). The Boost Clock was actually exceeding its limit. Is there a chance it's more likely that the thermal paste was poorly applied?

[bigbenisdaman]

Publicado originalmente por Iggy Wolf:

Ok, so what does it mean when your card is practically EXCEEDING the max temp limit? MSI Afterburner shows me that its temp limit is 83 C. It reached 86 C in-game, the fans started to run at 100% (the graph showing them attempting to run at 112%). The Boost Clock was actually exceeding its limit. Is there a chance it's more likely that the thermal paste was poorly applied?

The limit in afterburner isn't the max limit. You can set that lower and it just ramps the fans up to maintain temps. Once it hits 90C it'll start downclocking and not boost as high/at all.

Have you tested running with your side panel off yet? Even reviews though have it running in the upper 70's, so in your closed off case with limited airflow, it could very well be hotter.

[Iggy Wolf]

Publicado originalmente por bigbenisdaman:

Publicado originalmente por Iggy Wolf:

Ok, so what does it mean when your card is practically EXCEEDING the max temp limit? MSI Afterburner shows me that its temp limit is 83 C. It reached 86 C in-game, the fans started to run at 100% (the graph showing them attempting to run at 112%). The Boost Clock was actually exceeding its limit. Is there a chance it's more likely that the thermal paste was poorly applied?

The limit in afterburner isn't the max limit. You can set that lower and it just ramps the fans up to maintain temps. Once it hits 90C it'll start downclocking and not boost as high/at all.

Have you tested running with your side panel off yet? Even reviews though have it running in the upper 70's, so in your closed off case with limited airflow, it could very well be hotter.

I should clarify that apparently this occurred because I had clicked the "Auto" option under fan to be turned off, most likely therefore no longer using the custom fan curve I set. That said, the default factory fan speed is CLEARLY bad since on its own, without any adjustments by Afterburner, it starts to hit 86 C, and I notice stuttering in the game, with the FPS fluctuating between 120 and 80 FPS. The custom fan curve still lets it get to 81 C though. I wonder if I should just lower the temp limit (linked with GPU power limit) to 76, which will put the power limit at 90%. At least then, it'll reduce power consumption and temps while not affecting performance too greatly.

[[☥] - CJ -]

It was me, and the clean install NVIDIA option does nothing, all it does is reset the NVIDIA CP settings, nothing more, which isnt helpful.

run DDU in safe mode, select the NVIDIA driver, click clean and restart
install the latest RTX driver, see if anything changes

Also, in the NVIDIA CP i would turn off the Shader Cache.

Enable VSync in any games you play before you do any of this, see how the temps are then.

[Iggy Wolf]

So I decided to test it again. The temps constantly stay at 83 C despite the fan speed being at 100%. I really hope it's just a poor factory thermal paste job. My case may be pre-built and mid-tower but the motherboard is an microATX and there's barely much space occupied in the case, leaving plenty of room for ventilation. Hence, why my GTX 970 rarely exceeded 73 C. It was loud though during gaming, presumably with the fans spinning at 80-100%. But at least it was keeping the card cool. This one, even at max fan speed, still runs hot.