GTX 960 SC at 89C after 15 minutes of Furmark, could it be botched thermal paste?

[hsjj3]

I've got a EVGA GTX 960 SC. Ran Furmark for 15 minutes. Stock GPU settings, but a constant 45% fan speed (which is slightly higher than the default fan speed of 40%). I increased the temp limit to 95C but kept the power limit at 100%.

The temperature reading was 89C after 15 minutes, with ambient room temperature of around 29-30C.

I saw the Furmark values for the GTX 960 SC on various forums where people have overclocked their cards (note: mine is stock settings) and despite running at lower fan speeds their readings never broach 70C. It's always around 65C. I know one guy, whom I asked to run Furmark with a 45% fan speed at stock settings, and he reported a 61C max temperature.

So you see, 61C > 89C is nearly a 50% jump!

My previous GPU in this exact same PC, EVGA GTX 750 SC, hit a high of 68C compared to the 57C reported by review sites at the time for the same GPU model. It was a difference of 20% higher back then.

So is this a symptom of a botched thermal paste application? I mean sure, my case has poor ventilation and due to living on the equator I have higher ambient temps than most people, but still...I think a 28C difference for the exact same card running the exact same application (Furmark) is hard to digest.



How often can a botched thermal paste application occur at the factory?

 

[UsandThem]

It's extremely rare for a video card to not have paste from the factory.

It is more likely it is your case airflow. If your room temps are 86f, your case is probably well over 100f.

My room temperature is 75f, I have four 140mm case fans (Noctua on lower RPM), and my case temp while gaming hits around 96f. My GTX 970 reference card fan goes higher than 45% while gamiing.

To to see if your case is too hot, leave the side of the case off and use a fan to blow into it. If your video card temp drops significantly, it is your room temp and case temp causing the problem. Increase the fan speed of your card up from 45%, because your conditions will probably need a higher fan speed to keep temps down.

If not, you might just have a defective card, or maybe the cooler isn't making solid contact.

 

[96Firebird]

Why are you locking your fan speed at a constant rate rather than allowing it to go up with temperature?

 

[Despoiler]

You should not be using the Furmark power virus to test anything. There is no game that will ever come close to approaching the types loads it puts on a graphics card.

 

[Erithan13]

You're expecting far too much out of the cooler if you're limiting it to 45% speed while running furmark at 30*C ambient, that's practically begging for overheating/throttling/thermal shutdown. No need to abuse the card like this, if you use it for gaming, then the temperatures under gaming with an appropriate fan profile are what you want to worry about.

 

[sm625]

Are you sure it is actually 89C? Do you have one of those handy little temperature probe guns?

 

[hsjj3]

I know Furmark is worst case scenario, but I needed to do it to compare with benchmarks and forum posts. Regardless, even with running TW3 uncapped at the swamp area, my temps rise to 86C.

If I remove the side cover, the Furmark max temp drops to 77C, and TW3 drops to 76C.

This however is still hotter than the guy who got 61C at 45% on Furmark in 22C room temperature on the same GPU model as me.

I've got some Thermaltake TG-3. Hopefully it can reduce my maximum in-case Furmark temps by 10C.

 

[UsandThem]

I know Furmark is worst case scenario, but I needed to do it to compare with benchmarks and forum posts. Regardless, even with running TW3 uncapped at the swamp area, my temps rise to 86C.

If I remove the side cover, the Furmark max temp drops to 77C, and TW3 drops to 76C.

This however is still hotter than the guy who got 61C at 45% on Furmark in 22C room temperature on the same GPU model as me.

I've got some Thermaltake TG-3. Hopefully it can reduce my maximum in-case Furmark temps by 10C.

A 22c room compared to a 30c room is a huge difference. Factor in case cooling as well, while running a benchmark at fixed fan settings is the issue.

 

[Flapdrol1337]

Don't run furmark, it's silly. Use unigine heaven or something, it loops permanently too.

 

[Despoiler]

I know Furmark is worst case scenario, but I needed to do it to compare with benchmarks and forum posts. Regardless, even with running TW3 uncapped at the swamp area, my temps rise to 86C.

If I remove the side cover, the Furmark max temp drops to 77C, and TW3 drops to 76C.

This however is still hotter than the guy who got 61C at 45% on Furmark in 22C room temperature on the same GPU model as me.

I've got some Thermaltake TG-3. Hopefully it can reduce my maximum in-case Furmark temps by 10C.

It's not a worst case scenario. It's a no case scenario. You will never under any circumstance find a real world game or application that will load a card like a power virus does.