Question Nvidia RTX 2080 high temp ?

[Neocro]

Hello. I got used 2080 card and changed thermal paste but this temps seems 2 high for me
and that's like after 1 min of play
My fans sometimes goes over 3000 RPM and its very loud
Case is Panteks P600S with 5 fans

Suggestions ?

 

[Phaaze88]

Gpu memory and gpu hot spot look very good, but gpu core temperature is missing.

 

[Neocro]

Thats after 1 min play. If i play longer temp goes to 90 and RPM over 3000

 

[Phaaze88]

90C on... what?
The fans are likely reacting to the gpu core thermals more than the other 2, both of which have higher thermal thresholds.
Hot spot can go up to 110C.
Memory up to 105C.
Gpu core, up to 88C, but thermal throttling is preset to 83C.

 

[Neocro]

90C on... what?
The fans are likely reacting to the gpu core thermals more than the other 2, both of which have higher thermal thresholds.
Hot spot can go up to 110C.
Memory up to 105C.
Gpu core, up to 88C, but thermal throttling is preset to 83C.

I was reading on nvidia forum that max temp for that card should be 85 under 100% load. So this seems 2 much

 

[Phaaze88]

Gpu core temperature limits listed here: https://www.techpowerup.com/vgabios/203751/nvidia-rtx2080-8192-180828
Can't find the old thread that mentioned hot spot limit of 110C, so take it with a grain of salt, if you wish... though it's not unusual for that sensor to be reading to be 10-20C higher than gpu core.
Memory is here: https://www.micron.com/products/ultra-bandwidth-solutions/gddr6

Spikes aren't a major issue, but if it's sustaining 85C core, then that's a problem, because it means the thermal throttling function(drop clocks and voltage) isn't doing enough to get temperatures down. The card would likely need a repaste at that point.

 

[Neocro]

Gpu core temperature limits listed here: https://www.techpowerup.com/vgabios/203751/nvidia-rtx2080-8192-180828
Can't find the old thread that mentioned hot spot limit of 110C, so take it with a grain of salt, if you wish... though it's not unusual for that sensor to be reading to be 10-20C higher than gpu core.
Memory is here: https://www.micron.com/products/ultra-bandwidth-solutions/gddr6

Spikes aren't a major issue, but if it's sustaining 85C core, then that's a problem, because it means the thermal throttling function(drop clocks and voltage) isn't doing enough to get temperatures down. The card would likely need a repaste at that point.

Well that is problem. It has been in store for termal pada replace, thermal paste new and fan cleanup. Im gona try tomorow do it my self

 

[Phaaze88]

Well that is problem. It has been in store for termal pada replace, thermal paste new and fan cleanup. Im gona try tomorow do it my self

Can't use just any paste on the gpu die, like one can with cpus. Some of them are prone to getting pushed/pumped off.
Result: high gpu core and hot spot, or high hot spot; depends on how much got shoved off.

Thermal pads should be a last resort, as it can be a costly trial and error process. It's not enough to know the depth of the original pads, as they all come in different degrees of hardness.
Too firm, and the cooler cold plate no longer mates well with the gpu die. Too soft, and the memory doesn't make solid contact with the cooler's heatsink.
It's best left alone until there are memory thermal problems, or pads were torn.
Result: high gpu hot spot or memory.

 

[KananX]

90° on the hotspot aren’t that alarming rather normal for most coolers on high end cards. Remember this tool only showed you the hot spot, the normal temp is hidden and will be way lower, possibly under 80°.

Just use afterburner to monitor your temps.

 

[KyaraM]

Get hwinfo64 and check temps there. Just knowing min and max temps isn't very useful, they are only shortterm spikes. Besides, 90°C hotspot is nowhere near dangerous. Knowing averages is more important. Also which metrics you are even looking at.

 

[Neocro]

Get hwinfo64 and check temps there. Just knowing min and max temps isn't very useful, they are only shortterm spikes. Besides, 90°C hotspot is nowhere near dangerous. Knowing averages is more important. Also which metrics you are even looking at.

it was from HW info temp

I opened card today and thermal pads are toasted
Paste is toasted also so they only used air in service nothing else. In Monday I'm getting my money back from service since they done nothing

New temps after 20 min of Forza Horizon 5

 

[KyaraM]

it was from HW info temp

I opened card today and thermal pads are toasted
Paste is toasted also so they only used air in service nothing else. In Monday I'm getting my money back from service since they done nothing

New temps after 20 min of Forza Horizon 5

That's HWMonitor, not HWInfo. Different things. But sounds like you found the issue already. Still, HWInfo is netter imho.

 

[Neocro]

That's HWMonitor, not HWInfo. Different things. But sounds like you found the issue already. Still, HWInfo is netter imho.

it does same job i just downloaded and it shows same temps