Strange HotSpot readings from my RTX 2070 Super

[Lorx]

Hi there!

I'm having strange temps reading by HWinfo about my GPU HOTSPOT, basically when playing Apex or any "demanding" games i hit temps around 80°C on the GPU but 105°C (stable) on the hotspot!

After a game lasted 10 mins (for testing purposes) my highest reached temp are:

GPU: 78.5°C
GPU Hot Spot: 105.7°C

GPU MODEL IS: Gigabyte RTX 2070 Super Gaming OC 3x 8G
MOBO IS: MSI X570 TOMAHAWK
HWINFO VERSION: 7.26

For testing purposes i also set all my fan (GPU fans too) at maximum RPM and launched the game, still the same! GPU temps kinda nominal but hotspot wanting to reach 105°C.
I discovered, hovering the cursor on the GPU hotspot reading that the temp is "unconfirmed" like there is no real sensor for the hotspot cause the exact message says:

"this temperature in unconfirmed, but assumed to be the value from the hottest thermal sensor"

Having this new info, i tried a series of thing, i left my fans setting as i always had with custom curves for case fan, and not touching the GPU curves, i launched the game, again high hot spot temp as expected, but as soon as i TAB OUT on my second monitor (game still running full screen on my first monitor) hotspot reading drops INSTANTLY by 20-ish°C, as soon as i come back to the game the reading skyrockets with no apparent reason.

And now i'm here wondering if is just a fake or an exaggerated reading considering is an assumed value or if i need to be worried.

I did this post cause i'm not really comfortable with the idea of opening my GPU to apply new thermal paste and thermal pads, especially if it's not needed and it's a strange reading.

Thanks

 

[Martin]

I believe this value is real and reflecting a very high temperature of some other part of the GPU. It could indeed be a result of improperly applied thermal paste or cooler on GPU, memory chips or VRM PWMs.
Have you also observed performance throttling when the hotspot reaches such levels?

 

[Lorx]

I believe this value is real and reflecting a very high temperature of some other part of the GPU. It could indeed be a result of improperly applied thermal paste or cooler on GPU, memory chips or VRM PWMs.
Have you also observed performance throttling when the hotspot reaches such levels?

I'm doing other tests to confirm the thermal throttling but as of HWInfo tells me the thermal throttle happens. I'll try the route of undervolting first, then i may be do a repasting and applying new thermal pads.
What it looked strange to me was the SUDDEN drop of 20-ish degrees insantly when tabbing out. But i guess after almost 2 year of every day use the thermal compound just dried up and the pads are maybe failing.

 

[Kiriakos-GR]

I'm doing other tests to confirm the thermal throttling but as of HWInfo tells me the thermal throttle happens. I'll try the route of undervolting first, then i may be do a repasting and applying new thermal pads.
What it looked strange to me was the SUDDEN drop of 20-ish degrees insantly when tabbing out. But i guess after almost 2 year of every day use the thermal compound just dried up and the pads are maybe failing.

I would worry of the health of DC fan.
Thermal pads they are immortal, if they are factory installed do not touch them.
Thermal paste survives for 10 years.
If the card has factory installed paste and pads, then ask help from PC builder professional, so him to take apart the cooler and clean it.