Question Temp too high

Jun 13, 2021
44
1
35
0
Hi,

was gaming as usually on my PC
-ryzen 3900x
-2080ti
-noctua ndh cooler

When I saw the temperatures were too High.
I used HWinfo.

GPU temp under furmark stress test 73 C but the hotspot temp Is 85C

Ryzen temp under prime95 Is 73 C so pretty good.
But when I use furmark and prime95 together I get 75 for GPU and 83 for ryzen.
Also hw info says that my chipset temp Is 70C no matter what, but controllino with a thermocoupler the heatsink Is at 39C


Are these normal temps?
 

jasonf2

Honorable
Oct 11, 2015
444
118
11,040
40
85 degrees C is at the top of current gen chips normal temperature operating window. Modern system designs CPUs\GPUs all dynamically set their clock speed. I don't know how much you know about clocking but in the simplest sense the higher the clock the more heat produced (The reasons why are a little more complicated than that.). So when we benchmark or pull heavy loads the chips boost to the highest clocks they can. So on any assembly you are watching there is always going to be a hottest part, which in turn sets the throttle. If you were in excess of 85-90 I would be worried but this just sounds normal as something in there will always run at the top when benching or gaming. Notwithstanding secondary design limitations that is why over clockers are able to use Liquid N2 to get crazy high clocks. They are able to pull enough heat away to let the system run up to impedance limits (related to transistor design) not the heat transfer limits faced with ambient air temps.

If you want to try to keep the hotspot down the best and only real thing to do is make sure you have good case air flow to components. I would make sure there are no blockages and the card is clean (no dust bunnies).
 
Reactions: keith12
Jun 13, 2021
44
1
35
0
85 degrees C is at the top of current gen chips normal temperature operating window. Modern system designs CPUs\GPUs all dynamically set their clock speed. I don't know how much you know about clocking but in the simplest sense the higher the clock the more heat produced (The reasons why are a little more complicated than that.). So when we benchmark or pull heavy loads the chips boost to the highest clocks they can. So on any assembly you are watching there is always going to be a hottest part, which in turn sets the throttle. If you were in excess of 85-90 I would be worried but this just sounds normal as something in there will always run at the top when benching or gaming. Notwithstanding secondary design limitations that is why over clockers are able to use Liquid N2 to get crazy high clocks. They are able to pull enough heat away to let the system run up to impedance limits (related to transistor design) not the heat transfer limits faced with ambient air temps.

If you want to try to keep the hotspot down the best and only real thing to do is make sure you have good case air flow to components. I would make sure there are no blockages and the card is clean (no dust bunnies).
The max op temp of a 2080ti chip Is 85c
If the hotspot Is hotter Is dangerous?
 

Phaaze88

Titan
Ambassador
Furmark does have a use, as a Prime95(small FFT, All AVX off) for gpus, of which the purpose is testing gpu cooler capability.

I don't know why you'd run Prime95(settings unknown) and Furmark together, as that's far from realistic.

Hot spot temp will always be higher than gpu core, and that's it's ONLY 12C higher is a great thing.

The max op temp of a 2080ti chip Is 85c
Where did you get this from? The max temperature for the GPU DIE is 88C.
The hotspot doesn't follow the limit the die does.
 
Reactions: keith12
Jun 13, 2021
44
1
35
0
Furmark does have a use, as a Prime95(small FFT, All AVX off) for gpus, of which the purpose is testing gpu cooler capability.

I don't know why you'd run Prime95(settings unknown) and Furmark together, as that's far from realistic.

Hot spot temp will always be higher than gpu core, and that's it's ONLY 12C higher is a great thing.


Where did you get this from? The max temperature for the GPU DIE is 88C.
The hotspot doesn't follow the limit the die does.
Got from HWinfo
Runner together only to see how much the temp of the cpu Will affect the temp of the gpu
 

Phaaze88

Titan
Ambassador
Your cooling situation isn't as bad as you make it out to be - it really isn't.
IF your 2080Ti isn't a blower/turbo card, you could consider undervolting it to reduce some of it's power consumption - that would in turn reduce some of the energy that makes its way into the cpu cooler.
You could do the same with the 3900X, but I'm not that familiar with Ryzen; the most I could tell you is to turn PBO off and run Windows balanced power plan.

So 85c as hotspot Isnt an issue?
Negative.

Also, how can I control the fan Speed of the gpu without running afterburner in the background
That would require you to open the gpu, take the fan connectors and use an adapter like THIS to connect them to the motherboard. Then you could control them from bios.
 

jasonf2

Honorable
Oct 11, 2015
444
118
11,040
40
The max op temp of a 2080ti chip Is 85c
If the hotspot Is hotter Is dangerous?
If it goes high enough it will do damage to the chip. However more than likely the system is throttling to that temp so it will more than likely self regulate. Here is the technical reason that throttling reduces heat. Impedance by definition is the effective resistance of an electric circuit or component to alternating current, arising from the combined effects of ohmic resistance and reactance. So in high school physics what they taught you was probably only Ohms law which is that we have ohmic resistance (ohms), current flow (amps) and potential (voltage). In AC circuits (and pulsed DC) those three things exist but there are also some other things that are in play due to the fact the flow of current is reversing, or in our case starting and stopping (pulsed dc). So you have ohmic resistance, inductance (magnetic fields), capacitance (electric fields), inductive reactance and capacitive reactance. The reactance in a circuit is relative to frequency and really comes into play in high frequency circuits (ie your CPU or GPU running at gigahertz speeds). Switching frequency also effects transistor response. So without going into even more of a physics lesson as we change the operating frequency of a circuit the effective resistance of the same circuit changes and transistors behave differently. So in a cpu or gpu for all given purposes the lower the operating frequency the less voltage necessary to maintain stable switching. Just like in high school physics if you increase or decrease voltage you directly increase or decrease amperage so both are directly proportional to heat production. As all computer circuits are battling power consumption and heat production most modern systems try to dynamically keep voltage as low as possible by reducing frequency and core voltage to the minimum state needed. So under normal idle conditions most cards are running at an underclock, dynamically boost and only during intensive gameplay or computational activity will they ever really ramp up to the top of their thermal envelope, not necessarily frequency. So when you are benching you should expect the thermal envelope to be pushed somewhere in the card assuming that the chip is efficiently using its variable clock. If the top end of the envelope is 85 degrees by design, there is no damage being done by running here, your card is just maxed, which it should be on a benchmark.
 
Last edited:
Jun 13, 2021
44
1
35
0
If it goes high enough it will do damage to the chip. However more than likely the system is throttling to that temp so it will more than likely self regulate. Here is the technical reason that throttling reduces heat. Impedance by definition is the effective resistance of an electric circuit or component to alternating current, arising from the combined effects of ohmic resistance and reactance. So in high school physics what they taught you was probably only Ohms law which is that we have ohmic resistance (ohms), current flow (amps) and potential (voltage). In AC circuits (and pulsed DC) those three things exist but there are also some other things that are in play due to the fact the flow of current is reversing, or in our case starting and stopping (pulsed dc). So you have ohmic resistance, inductance (magnetic fields), capacitance (electric fields), inductive reactance and capacitive reactance. The reactance in a circuit is relative to frequency and really comes into play in high frequency circuits (ie your CPU or GPU running at gigihertz speeds). Switching frequency also effects transistor response. So without going into even more of a physics lesson as we change the operating frequency of a circuit the effective resistance of the same circuit changes and transistors behave differently. So in a cpu or gpu for all given purposes the lower the operating frequency the less voltage necessary to maintain stable switching. Just like in high school physics if you increase or decrease voltage you directly increase or decrease amperage so both are directly proportional to heat production. As all computer circuits are battling power consumption and heat production most modern systems try to dynamically keep voltage as low as possible by reducing frequency and core voltage to the minimum state needed. So under normal idle conditions most cards are running at an underclock, dynamically boost and only during intensive gameplay or computational activity will they ever really ramp up to the top of their thermal envelope, not necessarily frequency. So when you are benching you should expect the thermal envelope to be pushed somewhere in the card assuming that the chip is efficiently using its variable clock. If the top end of the envelope is 85 degrees by design, there is no damage being done by running here, your card is just maxed, which it should be on a benchmark.
Yea was Just trying to figure out if this chip will damage with an hotspot of 85C

Also as an RF tech I am kinda familiar with this concepts xD
Also good basic explanation
 
Jun 13, 2021
44
1
35
0
Your cooling situation isn't as bad as you make it out to be - it really isn't.
IF your 2080Ti isn't a blower/turbo card, you could consider undervolting it to reduce some of it's power consumption - that would in turn reduce some of the energy that makes its way into the cpu cooler.
You could do the same with the 3900X, but I'm not that familiar with Ryzen; the most I could tell you is to turn PBO off and run Windows balanced power plan.


Negative.


That would require you to open the gpu, take the fan connectors and use an adapter like THIS to connect them to the motherboard. Then you could control them from bios.
I thought that maybe there could a native program that stores the Curves in the GPU itself
 

ASK THE COMMUNITY