PC Specs:
MSI MAG B550M Mortar
AMD R5 5600 (stock)
2x8GB Kingston HyperX Fury RGB DDR4 3200MHZ
PSU XPG Core Reactor 650W
SSD NVME M.2 Kingston 960GB
SSD SATA SAMSUNG EVO 960GB
Gigabyte RTX 2080 OC
So, I built this pc in 2020 and I used to have a GTX 1650 super from Asus (tuf 2-fan model). It always have worked just fine, super silent, not much heat (I don't know exactly what my air cooler is, but my cpu never reaches 70ºC).
My brother bought a 4070 Ti, so his RTX 2080 was getting dusty in his shelf (and I believe he kept it for more than a year). He now decided to give me the card (he lives in Australia and I live in Brazil, so it was not as simply as just handing me over the gpu). I uninstalled the drivers from my 1650s and put the 2080 on, installed the drivers again, but when I tried to test it, it reached 90ºC+ almost instantly, it was something like 3 seconds, doing a CS2 benchmark. Initially, I thought it was a driver problem, but after reinstalling the driver thrice and restarting the pc over and over again, I assumed I'd have to open the card and change the thermal paste. This is the card opened:
As you can see, the thermal paste was very bad, and one of the thermal pads just fell off as I opened the card. I changed the thermal paste and it instantly decreased 10ºC of temperature, remaining at 78~81ºC at 100% usage.
This is not the strange part, as I was assuming that it would get hot. The strange part is that, the maximum fan RPM I can achieve with MSI Afterburner is like 3300 RPM, but the card gets 4400+ RPM when it reaches those 80ºC. In the FurMark test, I can see that the hotspot is about 107ºC, as the GPU is at 80ºC. It reaches 80ºC in like less than 10 seconds, and that's not all.
So, the GPU is getting SUPER noisy, and it really doesn't have to be at 100% usage. This is why the title is "awkwardly overheating". I was playing Black Ops III, with unlocked framerate. Getting about 120 frames at FullHD ultra. The GPU was obviously at 100% usage, and getting that noisy 4400 RPM fan rotation. The REALLY awkward part is that, when I locked the framerate at 75fps, the card was at 50% usage, and it was still getting the exact same temperatures, same fan speed, same hotspot temperature. How could the card be getting 80ºC at both 50% and 100% usage? How could it be getting the same fanspeed and maintaining the same temperature, not mattering if it is at 50% or 100%?
Can someone help me with this? What should I do? Undervolt? Did I applied the thermal paste wrongly?
Right now where I live is 9ºC ambient temperature, and the card gets 80ºC either way. My pc has a good airflow (that's why the CPU never gets too hot). While playing Warzone, I actually tried to put my ventilator pointed at my opened pc case, the CPU was at 42ºC and the GPU was still at those noisy 80ºC.
I changed the fan curve so it reaches 100% fanspeed when it gets to 70ºC. That helped delaying the overheating in a few seconds, but it overheats anyway. Before applying thermal paste, I was getting a lot of thermal throttling playing CS2, clocks dropping to 300Mhz, unplayable. After the thermal paste, I was not feeling substential drops, and I still don't. I've tested in CS2, Bo3, Warzone, and I can play normally, not having to worry with anything but the super noisy gpu. Is this dangerous? I believe it's not good for the GPU's wear to be reaching 100ºC+ (hotspot) and then getting down to 40ºC, just so it can reach 100ºC again within a few seconds.
I've tried some undervolts, these were the results (FurMark 1080p benchmark):
8045 points - standard mv
7177 points - 950 mv
6920 points - 925 mv
7048 points - 900mv
6743 points - 875mv
7103 points - 850mv
7022 points - 825mv
6961 points - 800mv
7061 points - 775mv
7442 points - 750mv
7557 points - 725mv
Here's the prints (2) I took playing BO3:
I know that the GPU FAN says "40%", but it was just bugged, trust me, the card was noisy as hell the entire playthrough. The temperature of the chip and the hotspot remained the same despite undervolting.
MSI MAG B550M Mortar
AMD R5 5600 (stock)
2x8GB Kingston HyperX Fury RGB DDR4 3200MHZ
PSU XPG Core Reactor 650W
SSD NVME M.2 Kingston 960GB
SSD SATA SAMSUNG EVO 960GB
Gigabyte RTX 2080 OC
So, I built this pc in 2020 and I used to have a GTX 1650 super from Asus (tuf 2-fan model). It always have worked just fine, super silent, not much heat (I don't know exactly what my air cooler is, but my cpu never reaches 70ºC).
My brother bought a 4070 Ti, so his RTX 2080 was getting dusty in his shelf (and I believe he kept it for more than a year). He now decided to give me the card (he lives in Australia and I live in Brazil, so it was not as simply as just handing me over the gpu). I uninstalled the drivers from my 1650s and put the 2080 on, installed the drivers again, but when I tried to test it, it reached 90ºC+ almost instantly, it was something like 3 seconds, doing a CS2 benchmark. Initially, I thought it was a driver problem, but after reinstalling the driver thrice and restarting the pc over and over again, I assumed I'd have to open the card and change the thermal paste. This is the card opened:
As you can see, the thermal paste was very bad, and one of the thermal pads just fell off as I opened the card. I changed the thermal paste and it instantly decreased 10ºC of temperature, remaining at 78~81ºC at 100% usage.
This is not the strange part, as I was assuming that it would get hot. The strange part is that, the maximum fan RPM I can achieve with MSI Afterburner is like 3300 RPM, but the card gets 4400+ RPM when it reaches those 80ºC. In the FurMark test, I can see that the hotspot is about 107ºC, as the GPU is at 80ºC. It reaches 80ºC in like less than 10 seconds, and that's not all.
So, the GPU is getting SUPER noisy, and it really doesn't have to be at 100% usage. This is why the title is "awkwardly overheating". I was playing Black Ops III, with unlocked framerate. Getting about 120 frames at FullHD ultra. The GPU was obviously at 100% usage, and getting that noisy 4400 RPM fan rotation. The REALLY awkward part is that, when I locked the framerate at 75fps, the card was at 50% usage, and it was still getting the exact same temperatures, same fan speed, same hotspot temperature. How could the card be getting 80ºC at both 50% and 100% usage? How could it be getting the same fanspeed and maintaining the same temperature, not mattering if it is at 50% or 100%?
Can someone help me with this? What should I do? Undervolt? Did I applied the thermal paste wrongly?
Right now where I live is 9ºC ambient temperature, and the card gets 80ºC either way. My pc has a good airflow (that's why the CPU never gets too hot). While playing Warzone, I actually tried to put my ventilator pointed at my opened pc case, the CPU was at 42ºC and the GPU was still at those noisy 80ºC.
I changed the fan curve so it reaches 100% fanspeed when it gets to 70ºC. That helped delaying the overheating in a few seconds, but it overheats anyway. Before applying thermal paste, I was getting a lot of thermal throttling playing CS2, clocks dropping to 300Mhz, unplayable. After the thermal paste, I was not feeling substential drops, and I still don't. I've tested in CS2, Bo3, Warzone, and I can play normally, not having to worry with anything but the super noisy gpu. Is this dangerous? I believe it's not good for the GPU's wear to be reaching 100ºC+ (hotspot) and then getting down to 40ºC, just so it can reach 100ºC again within a few seconds.
I've tried some undervolts, these were the results (FurMark 1080p benchmark):
8045 points - standard mv
7177 points - 950 mv
6920 points - 925 mv
7048 points - 900mv
6743 points - 875mv
7103 points - 850mv
7022 points - 825mv
6961 points - 800mv
7061 points - 775mv
7442 points - 750mv
7557 points - 725mv
Here's the prints (2) I took playing BO3:
I know that the GPU FAN says "40%", but it was just bugged, trust me, the card was noisy as hell the entire playthrough. The temperature of the chip and the hotspot remained the same despite undervolting.
Last edited: