Does Thermal Paste have an impact on in-game FPS/Stutters?

NocturiusHD · 14 May 2020 at 09:38

Hello forum,

So recently i changed my Thermal Paste on my GTX 1080 GPU with Grizzly kryonaut. I have read tons of posts about this topic, but only posts about "bad temperatures and overheating", which is not my concern since everything is super cool due to AIO cooling (56-60 hot under full load). Btw i have i7-9700K CPU, 16 GB 3200 Ram and 650W PSU to support my GPU.

I have done this procces before and followed JayTwoCents guide again, and used same amount of Thermal Paste as him.

View: https://www.youtube.com/watch?v=6HbCY3-tun0&t=0s

But.. After applying a new layer of Thermal Paste i have noticed a slighety decrease in FPS and some game stuttering, but overall better GPU temperature. So i was wondering if Thermal Paste on GPU has a direct impact on gaming performance (FPS/STUTTERS)? Im not afraid that i have appylied to much Thermal Paste, but not sure if to much can cause fps drops and stutters while gaming.

I know that i could just change the GPUs Thermal Paste, but this would be a pain in the ass since i am happy with the current GPU temperature. Am i overthinking this or is this a known issue? P.S. I have PC OCD haha

benjii · 14 May 2020 at 12:29

The only way thermal paste can affect FPS is if it were improperly applied and the GPU was over heating and throttling. As you've said the temperatures are better than they were previous, then what I said isn't the case. The FPS problems are unrelated.

NocturiusHD · 14 May 2020 at 12:41

benjii said:
The only way thermal paste can affect FPS is if it were improperly applied and the GPU was over heating and throttling. As you've said the temperatures are better than they were previous, then what I said isn't the case. The FPS problems are unrelated.

Thanks for the anwser Sir

merlin6r · 15 May 2020 at 01:12

Is this in several different games ?

NocturiusHD · 15 May 2020 at 06:51

merlin6r said:
Is this in several different games ?

Yes it did. I repasted the GPU with thin spread method and only tightened the screws with my fingers. This solved my problem. Before i had tightened the screws to much i think. So the Thermal Paste was spread "to much".

merlin6r · 15 May 2020 at 08:44

NocturiusHD said:
Yes it did. I repasted the GPU with thin spread method and only tightened the screws with my fingers. This solved my problem. Before i had tightened the screws to much i think. So the Thermal Paste was spread "to much".

Glad to hear you sorted it!

doyll · 15 May 2020 at 13:57

NocturiusHD said:
Yes it did. I repasted the GPU with thin spread method and only tightened the screws with my fingers. This solved my problem. Before i had tightened the screws to much i think. So the Thermal Paste was spread "to much".

Probably over-tightening was distorting surfaces and this resulted in poor surface mating / TIM print. TIM is not supposed to be a layer between chip and cooler. TIM is used to fill microscopic voids in surface metal and areas where metal does not contact metal only. TIM has a heat transfer rate of 3-15W/m-K while copper is 400 W/m-K, aluminium is about 200W/m-K and air is only 0.024W/m-K. So while TIM is much better at transferring heat than nothing, it is not near as good as direct metal contact with TIM only filling those microscopic voids.

merlin6r · 16 May 2020 at 03:39

doyll said:
Probably over-tightening was distorting surfaces and this resulted in poor surface mating / TIM print. TIM is not supposed to be a layer between chip and cooler. TIM is used to fill microscopic voids in surface metal and areas where metal does not contact metal only. TIM has a heat transfer rate of 3-15W/m-K while copper is 400 W/m-K, aluminium is about 200W/m-K and air is only 0.024W/m-K. So while TIM is much better at transferring heat than nothing, it is not near as good as direct metal contact with TIM only filling those microscopic voids.

Yeah definitely, I think a lot of new builders think that the thermal paste is actually something that helps transfer the heat, rather than it being just something that fills the gaps and is better than air in between the surfaces.

I aim for a layer about as thick as a coat of paint. I apply the paste, spread it out a bit and then swipe (pardon the pun) over the top with an old credit card. The cpu looks like it's just been painted grey. [I hope] this results in some areas getting full metal to metal contact and the surplus squishing into the recesses...

benjii · 16 May 2020 at 09:41

Too much thermal paste is a myth. The only negative effect is too little. You cannot end up with too much because the amount of pressure applied by the cooler mounting squeezes excess out.

This is one of many articles/videos where people have tested too much thermal paste.

https://www.gamersnexus.net/guides/3346-thermal-paste-application-benchmark-too-much-thermal-paste

doyll · 16 May 2020 at 12:58

benjii said:
Too much thermal paste is a myth. The only negative effect is too little. You cannot end up with too much because the amount of pressure applied by the cooler mounting squeezes excess out.

This is one of many articles/videos where people have tested too much thermal paste.

https://www.gamersnexus.net/guides/3346-thermal-paste-application-benchmark-too-much-thermal-paste

If thermalpaste is conductive (several are) then using too much it will cause it to sqush out onto motherboard shorting out and destroying it. Most of use consider that a negative effect.

If thermalpaste is thin enough to push out and mount has enough pressure topush out surplus so final seat is metal to metal with TIM filling voids end result will be the same. If thermalpaste is thick and mounting pressure is not high enough and TIM does not flow easy there will be a layer between base and IHS with resulting temps being warmer. We are talking a few degrees, similar to the differences between top TIMs.

It appears Thermal Grizzly Kryonaut rated 12.5 W/m-K is thin enough for excess to flow out. But if there are any air bubbles entrapped in the seat it will pump TIM out and temps will rise. There have been many reports of Thermal Grizzly having higher temps after awhile. OcUK has had to re-seat several delidded CPUs they had used Thermal Grizzly on. Not sure why, maybe it's pumping out, maybe drying up.

Years ago Rafael Coetho of Hardware Secrets tested different amounts of Arctic Silver Céramique under a Zalman CNPS9900 MAX on GA-P55A-UD6 with i7 860 CPU and found different amounts /applications had 3c difference. Not a huge difference, but definitely a difference.

https://www.hardwaresecrets.com/what-is-the-best-way-to-apply-thermal-grease-part-1/9/