Question GPU Hotspot Temperature 105C during Gaming (GIGABYTE RTX 2070 SUPER WINDFORCE OC 3X 8G)

[Beletov]

Hello,

I have had this graphics card for 2 years and 4 months. I have recently noticed that the GPU Hot Spot Temperature measured on the card (HWINFO) is over 100C, sometimes going as high as 105C and the average hovering around 100C during a gaming session. The card is a GIGABYTE 2070 SUPER WINDFORCE OC 3X 8G. The card is NOT overclocked and was never overclocked, it is running on factory settings. The internals of the card were never tampered with. I have attached a screenshot of HWINFO displaying the maximum Hotspot Temperature being 105C while the maximum normal GPU Temperature was 75C. I would like to know if I should be worried, is my card dying, or am I only going to experience lower FPS as a result of the high Hotspot Temperature? What is the long-term result of my GPU running at high Hotspot Temperatures?

Thank you in advance,

Beletov

 

[Phaaze88]

I would like to know if I should be worried, is my card dying, or am I only going to experience lower FPS as a result of the high Hotspot Temperature?

Worried? No.
Dying? No.
Lower fps? No. Hot spot doesn't throttle until at least 110C.

Either you, a friend you trust, or an associate at a shop: clean off the old thermal paste on the gpu die, and apply new.
If not taking it to a shop: Arctic MX-4, Noctua NT-H2, CM Mastergel Pro V2, Gelid GC-Extreme are a few suggestions for pastes.

What is the long-term result of my GPU running at high Hotspot Temperatures?

None. It's just one of numerous sensors on the gpu die.
Gpu temperature is based on an average.

The card is NOT overclocked and was never overclocked, it is running on factory settings.

FYI, It IS overclocked. It's just automatically done for you via the built in boost algorithm, and Gigabyte applied a small core clock bump on top of it.

 

[Beletov]

Worried? No.
Dying? No.
Lower fps? No. Hot spot doesn't throttle until at least 110C.

Either you, a friend you trust, or an associate at a shop: clean off the old thermal paste on the gpu die, and apply new.
If not taking it to a shop: Arctic MX-4, Noctua NT-H2, CM Mastergel Pro V2, Gelid GC-Extreme are a few suggestions for pastes.


None. It's just one of numerous sensors on the gpu die.
Gpu temperature is based on an average.




FYI, It IS overclocked. It's just automatically done for you via the built in boost algorithm, and Gigabyte applied a small core clock bump on top of it.

Thank you for the very informative reply. Just as a follow up, should I be messing with the internals while it is still in warranty?

 

[Phaaze88]

Thank you for the very informative reply. Just as a follow up, should I be messing with the internals while it is still in warranty?

It's ultimately your choice to make... but if you're not comfortable opening it up, don't do it, and let the company(Gigabyte) handle it.

 

[Beletov]

It's ultimately your choice to make... but if you're not comfortable opening it up, don't do it, and let the company(Gigabyte) handle it.

Once again, thank you for your reply. I am in quite a weird situation, GIGABYTE does not officially operate in my country and I am not comfortable with working with the internals because the card is still in the warranty period. Will there be any significant negative effects, if my GPU runs at that hotspot temperature in the future? Is changing the thermal paste necessary? I am completely fine with the GPU being slightly louder as a result of the higher hotspot temperatures. I have also contacted GIGABYTE regarding this matter but as it stands my previous experiences with GIGABYTE support were very poor so I do not expect them to actually help me in a meaningful way.

 

[Phaaze88]

What about a friend, or a shop you can take the gpu to? No dice on either?

Will there be any significant negative effects, if my GPU runs at that hotspot temperature in the future?

No, that's what the throttle protections are there for. Also, temperature spikes are not an issue. If it's sitting there - which the screenshot doesn't show - then it's a problem.

Is changing the thermal paste necessary?

If you desire to lower the hot spot reading, then yes.

 

[Beletov]

Thank you very much for your replies.

What about a friend, or a shop you can take the gpu to? No dice on either?

I will ask around and see what my options are.

No, that's what the throttle protections are there for. Also, temperature spikes are not an issue. If it's sitting there - which the screenshot doesn't show - then it's a problem.

The Hot spot temperature is not sitting at 105C rather, the average during a higher load is around 94C.

 

[Phaaze88]

You're welcome.

The Hot spot temperature is not sitting at 105C rather, the average during a higher load is around 94C.

That is fine.

 

[ELISSE]

What about a friend, or a shop you can take the gpu to? No dice on either?



No, that's what the throttle protections are there for. Also, temperature spikes are not an issue. If it's sitting there - which the screenshot doesn't show - then it's a problem.


If you desire to lower the hot spot reading, then yes.

i also have a technical question about this. what would happen if he never repasted and used the graphics card with 100C+ hot spot temps for a couple years? would there be any major fall off or significant drops in performance?

 

[Phaaze88]

i also have a technical question about this. what would happen if he never repasted and used the graphics card with 100C+ hot spot temps for a couple years? would there be any major fall off or significant drops in performance?

Huh? This is the same question, but with worded differently. Nothing really changes.
Plus, it should be repasted it if kept long enough.


The gpu does not respond to hot spot reading in the same manner as the gpu core one.
Hot spot: does nothing until the limit.
Gpu core: can change boost clock bins from a few degrees change.