Idle GPU in loop heat?

[Shadowman]

I was just curious if a gpu is idle and has under clocked itself because it's not being stressed.

 

But if your only stressing your cpu doing a full load test will the cpu temps be higher because of having a gpu in the loop as well even if it's sitting idle? 

[.Apex.]

1 minute ago, Shadowman said:

I was just curious if a gpu is idle and has under clocked itself because it's not being stressed.

It will under clock itself at idle yes, unless you force it

1 minute ago, Shadowman said:

But if your only stressing your cpu doing a full load test will the cpu temps be higher because of having a gpu in the loop as well even if it's sitting idle? 

Yes, every component in the loop will affect the overall temperature, but doesn't mean that it would be of significance, an idling GPU is generally less than 30W? something like that.

[Shadowman]

4 minutes ago, _Syn_ said:

It will under clock itself at idle yes, unless you force it

Yes, every component in the loop will affect the overall temperature, but doesn't mean that it would be of significance, an idling GPU is generally less than 30W? something like that.

So totally removing the gpu from the loop are we talking like 1c to 3c drop in temps for the cpu at full load? I was thinking more like 5c+

[.Apex.]

1 minute ago, Shadowman said:

So totally removing the gpu from the loop are we talking like 1c to 3c drop in temps for the cpu at full load? I was thinking more like 5c+

It would definitely not be 5C+, I'm thinking more of 1-2C at best, 30W is not difficult to dissipate but if you increase your CPU power by 30W for example that would have much more impact on the CPU temp because the heat is concentrated and is not dissipated as efficiently, but it won't affect the loop temp that much.

We're assuming the worst case scenario here, you haven't mentioned your GPU, as an example the RTX 2080 idles at only 11W, older GPU's might idle higher.

[Shadowman]

6 minutes ago, _Syn_ said:

It would definitely not be 5C+, I'm thinking more of 1-2C at best, 30W is not difficult to dissipate but if you increase your CPU power by 30W for example that would have much more impact on the CPU temp because the heat is concentrated and is not dissipated as efficiently, but it won't affect the loop temp that much.

We're assuming the worst case scenario here, you haven't mentioned your GPU, as an example the RTX 2080 idles at only 11W, older GPU's might idle higher.

It’s a RTX 2080ti

[.Apex.]

1 minute ago, Shadowman said:

It’s a RTX 2080ti

The RTX 2080 Ti idles a bit higher at 17W but the worst case scenario still holds true, assuming the GPU is truly idle and it's not being used

[.Apex.]

13 minutes ago, Shadowman said:

It’s a RTX 2080ti

You could also test that yourself, do a prime95 stress test for 10 mins then start a GPU stress test like Furmark at the same time and see how much that affects your CPU temps, that's a full 280W+ being put in a loop.

[nick name]

@Shadowman After you're done with your testing will you come back and tell us the temp differences with and without the GPU in the loop?  I'm curious.  

[.Apex.]

8 minutes ago, _Syn_ said:

You could also test that yourself, do a prime95 stress test for 10 mins then start a GPU stress test like Furmark at the same time and see how much that affects your CPU temps, that's a full 280W+ being put in a loop.

4:55

This is a CPU only stress test being done on a 6900K overclocked to 4.35Ghz (140W TDP Stock), so I'm assuming it's more in the range of 200W, as you can see at the end of 10 minutes, the loop only heated by 4C because the GPU is idle and the loop temp is the only thing affecting its temperature.

[Mick Naughty]

Whats the point exactly? Or is this is a too small for the system cooling setup?