HD6970 runs hotter than the internet!

[PSenapus]

Hello,

 

I've recently got an HD6970 reference design, made by Gigabyte.

I was excited, I got it for "cheap" so I set it up inside my HAF 922 case and fired up my heavily modded Skyrim and holy balls it gets hot.

 

I'm talking about 90 to 95 degrees with a almost stable 99% GPU load. These temperatures are exactly the same I get by running furmark. These temperatures seem to be above the avarage of people complaining on the internet about the same thing.

The graphics card is clean I sprayed lots of compressed air, the case has 1x 200mm fan intake on the front, 1x 230mm fan intake on the side, 1x 200mm exhaust at the top and 1x 120mm exhaust fan at the bottom. Ambient temperature is 23ish centigrade.

I'm also proud of my camble management work.

 

What do you think about these temperatures?
Should I worry? I'm not used to see such numbers so I'm quite scared

 

Let me know what you think and any suggestion will be great.

[creativename]

Its just that the 6970 is a hot card. Nothing you can really do about it except get an aftermarket cooler. AMD's reference coolers always sucked.

[coen113]

pictures? can you check if the fans are spinning properly? do the fans make a lot of noise?

[ERSCreeper]

Change the thermal paste. It's properly very old and has dried up.

[flibberdipper]

Make sure the fans are ramping up according to the temperature that the GPU is running at. Then check the thermal paste.

[lani100]

Check if the fan is spinning and i would suggest you replace the thermalpaste.

[SkilledRebuilds]

6970 is the DualGPU of the 6000 Series right?

 

Seems high, but expected if thats the case.

[helping]

6970 is the DualGPU of the 6000 Series right?

 

Seems high, but expected if thats the case.

no, single.

[CCap]

6970 is the DualGPU of the 6000 Series right?

 

Seems high, but expected if thats the case.

The 6990 was the Dual GPU of the 6000 series. 

[Humbug]

Is it clean or dusty?

[JR1911]

I had a pair of reference 6970s in Crossfire and they ran at about 92C when under load. I had them for almost three years without any problems so I wouldn't worry about the temps unless you experience crashing or other problems.

[LawrenceBarnes2013]

it's an AMD reference Cooler, you should expect that, just slap an aftermarket one on it . 

[PSenapus]

I had a pair of reference 6970s in Crossfire and they ran at about 92C when under load. I had them for almost three years without any problems so I wouldn't worry about the temps unless you experience crashing or other problems.

 

Damn so it isn't just mine that runs this hot...

Thanks for the information brother!

Is it clean or dusty?

 

F'N clean!

pictures? can you check if the fans are spinning properly? do the fans make a lot of noise?

The fan does spin, it spins also spins faster as soon as the temperature reaches 91 degrees, even faster at 93 degrees, then when the remperature goes down to 89 degrees it again spins slower. This cycle keeps repeating and imo makes little sense...  

 

I wouldn't like to check the thermal paste, I've had done this in the past with another reference card and the thermal pads on the ram and vrm were broken and I couldn't re-seat the heatsink properly on them without having to change the pads (and I can't easily find new pads here).

I've been told that vrm and ram do not make direct contact with the heatsink and they need "thick" pads in order to dissipate the heat, if this information is correct I can't just put some mx-2 instead of the pads...

[JR1911]

 

The fan does spin, it spins also spins faster as soon as the temperature reaches 91 degrees, even faster at 93 degrees, then when the remperature goes down to 89 degrees it again spins slower. This cycle keeps repeating and imo makes little sense...  

 

Yeah, the fan profile on the 6970 seems to be a bit weird. For me it would be almost silent most of the time, then it would suddenly spin up for a while and then go back to spinning slower again. It seemed to have only two settings, slow and full blast.

[PSenapus]

UPDATE:

I've set a custom fan profile with OC Guru II so that the fan would spin at 30% up to 70 degrees, then 70% at 70 degrees and 95% at 80+ degrees.

I would like to confirm that I'm glad reference coolers have been dropped by AMD.

The fan is fucking noisy, and maybe I should have caps locked the previous phrase just to give an idea, but at that speed the cooler does a good job at keeping the temperatures around 75-76 degrees. It rarely goes above 79 degrees even with a constant 99% GPU load.

 

UPDATE 2:
I also have received a reply from gigabyte's costumer support and all they told me is that the temperatures are high, I should add more fans and I shouldn't overclock the card.
Meaning they haven't even bothered reading my ticket, in which I have explained my system configuration (Case, fans and gpu clock).
Am I the only one here who dislikes Gigabyte?

[PSenapus]

Today I did another test:
Disabled the custom fan curve and started playing.
After 30 minutes of Skyrim I checked GPU Shark and the maximum temperature has been of 97 degrees.

 

That is high and the ambient temperature is also relatively low (I don't live in Canada and it's going to become really warm soon, and that worries me),

The fan spins at a maximum of 60% of the speed at such temperature.

I'm tempted to open it up and apply some mx-2.
What do you think?

I've never opened a reference cooler before and I'm scared of losing a screw or 2.
Halp.

[poseidon2988]

Change the thermal paste. It's properly very old and has dried up.

sounds like you should remove the old thermal paste yea:) use some good thermal paste:P

[nj4ck]

Thermal pads should be fine as long as you use a twisting moton rather than a pulling one to pull the cooler off. Use collaboratory liquid metal TIM and you could see a significant temp drop.

[PSenapus]

Hello,

Thank you for the replies.

I've opened the shroud and changed thermal compound.
I was scared because there was A LOT of thermal paste applied, it leaked all over the "things" around the die.
I've also noticed that 4 out of 8 RAM thermal pads were broken and the VRM pad was split in half too.
Applied some Arctic MX-2 I already had home and closed it.

Temperatures are pretty much the same.
2 degrees lower at load.

GPU-Z says that RAM and VRM are low temperature, the problem lies with the GPU reading number 1 that has reached a maximum of 95 degrees celsius in 10 minutes of Furmark
The GPU reading 2 is 7 degrees lower temperature and the GPU reading 3 has reached a maximum of 80 degrees.
Now unfortunately I have to leave this house and the GPU will remain in this computer (not mine), I wont be able to further play around with the cooler, however I would like to know if these readings are normal.
Why is GPU reading 1 so much higher than GPU reading 3?

[papeles]

those temps are normal for that gpu. It runs really hot unless you have an aftermarket cooler. As said before, the only think you can do is buy a new cooler

[RDilux]

i say remove the heatsink and get some GC Gelid Extreme on the chip

it might that the GPU thermal paste is become to old and hard so it won't make good contact with the surface 

[sgtsmokey]

I sold mine for £10, wasn't mine was it lol?

It was one of the reasons I upgraded.

An Arctic twin turbo fan for £30 should do it.