Witcher 3 testing for temps

[don_svetlio]

 So, I decided to Run around in Novigrad for about 20-30 minutes to see how my temps were.

Before, I'd get about 68*C on the GPU die and 66*C on the VRMs. Now that was before I reversed the airflow and ending up with negative case air pressure, before manually stopping my GPU fans and before using furmark as a test and before updating drivers. I updated them and the die went to about 66*C for the most part and the VRMs to about 65*C.

GPU utilization is 92-ish % and the average fps in that area is 52 with mostly low settings and V-sync on.

Temps now are 67*C on GPU die max from what I saw and 68*C on the VRMs. My question is, did I cause wear and tear via furmark, could the fan above them be spinning slower or is it just a case of that fan reusing warm air now (before it had 2 intakes blowing air to it, now they are exhausting warm air).

I looked at the fans while spinning and I think my eyes are playing tricks on me since if one were really spinning lower, then the die temps should also be higher as it's cooling mostly the actual heatsink which doesn't touch the VRMs.

Question is, can reversing the case fans have such an impact?

EDIT: RPMs displayed by Aida64 are the same as before mostly but idk if that applies to both fans. It should but idk.

[don_svetlio]

I should also mention that temps rose by about 2 degrees when the fan swap happened and that was before me touching anything. Maybe it's just the different airflow direction cuasing the air to get slightly more re-used?

[don_svetlio]

So I am tripleposting but since nobody is reading no harm done eh?

At any rate, tested again for 10 mins but in full screen rather than borderless window and after about 10-15 mins I ended up with 65*C on both while running around in that area. Probably the reversing of the case fans is causing the vrms to be slightly warmer. At any rate tho, under 70*C everything is more than fine from what I've heard so I will probably mark this solved

[SauronsV]

Your complaiming about hardly any difference

Smh

[Megahurt]

Testing temps with v-sync on ...

[don_svetlio]

Testing temps with v-sync on ...

Hmmm? considering I got about 52 fps and was at 92-95% usage I'd say that's a good way to test.

[Megahurt]

Hmmm? considering I got about 52 fps and was at 92-95% usage I'd say that's a good way to test.

vsync will make the load vary, and the temperature with it, considering you're only seeing 1°C difference, it's well within the margin of error that v-sync would introduce, depending on how demanding the scene is at each test.

[don_svetlio]

vsync will make the load vary, and the temperature with it, considering you're only seeing 1°C difference, it's well within the margin of error that v-sync would introduce, depending on how demanding the scene is at each test.

Welp, I was running around in the most taxing part of the game I could find (CPU usage fluctuates from 60-80%) and yeah, when outside of the city I get a constant 60 fps whereas in there it varies from 45-55 ususally so I do see what you mean. I hope I haven't damaged the VRMs with furmark or damaged the fans by stopping them. From what I was told, if it works as it is atm, it's probably not damaged but I'm paranoid

[Megahurt]

Welp, I was running around in the most taxing part of the game I could find (CPU usage fluctuates from 60-80%) and yeah, when outside of the city I get a constant 60 fps whereas in there it varies from 45-55 ususally so I do see what you mean. I hope I haven't damaged the VRMs with furmark or damaged the fans by stopping them. From what I was told, if it works as it is atm, it's probably not damaged but I'm paranoid

Today's gpus are limited by power limit, if you use a program that uses more power, they'll simply lower clockspeed and voltage, it won't allow it to cause any damage, don't worry about it. For example I get the same temps using furmark then I get from other benchmarks or games, the gpu just doesn't run at the same clockspeed.

[don_svetlio]

Today's gpus are limited by power limit, if you use a program that uses more power, they'll simply lower clockspeed and voltage, it won't allow it to cause any damage, don't worry about it. For example I get the same temps using furmark then I get from other benchmarks or games, the gpu just doesn't run at the same clockspeed.

So you're saying that regardless of stressing with Furmark or me stopping the brushless fans for a few seconds, it's highly unlikely to have caused any damage and what I'm basically seeing is fluctuating load + the formerly intake fans become exhaust thus pulling warmer air through the VRMs rather than pushing in cold air?

[Megahurt]

So you're saying that regardless of stressing with Furmark or me stopping the brushless fans for a few seconds, it's highly unlikely to have caused any damage and what I'm basically seeing is fluctuating load + the formerly intake fans become exhaust thus pulling warmer air through the VRMs rather than pushing in cold air?

Yeah to damage a gpu with load is difficult unless you modify the bios to remove the limitations of the card, and changing airflow direction could make the gpu hotter since you might be guetting hot air from the cpu cooler.

[don_svetlio]

Yeah to damage a gpu with load is difficult unless you modify the bios to remove the limitations of the card, and changing airflow direction could make the gpu hotter since you might be guetting hot air from the cpu cooler.

Yeah, the freezer sits around 40-45*C usually so that might be a factor. + it went from 1 intake to 2 exhaust. I guess if the 3Dmark Fire strike score went up after driver update by about 150, then I guess I haven't done any harm to the power delivery and it would seem the fans are okay. Still 68-70*C is good as for temps IIRC.