GPU Temps for 5870M (R2)

I recently decided to get back into Skyrim and Cysis 2 and noticed that the temps on my video cards are high - one card usually stays at about 82°C, and the other stays around 95°C with peaks of 101.5°C.

I have the laptop on a platform with additional fans positioned under the laptop's exhaust fans to draw heat away as it's dissipated.

Are these temperatures abnormally high? Dell replaced the motherboard a little while back and I'm concerned that heat sink might not be seated properly, and I'd like to avoid heat damage.

Thanks in advance for your thoughts!

 

I would say they are on the higher end for sure. These cards are not OVed or OCed right? When was the last time a repaste was done or when did Dell replace MOBO. To further help you along, i would use a higher grade paste such as Artic MX-4 if you feel comfortable reseating/repasting your cards yourself and clean the fins on the heatsink.

 

Thanks for your reply! The 100°C+ happens at the default voltage with standard clock speeds (700MHz / 1000MHz). I noticed earlier tonight that it's getting up to 104°C when the GPU load is only 60%. The motherboard replacement was about a year ago, but I haven't used it for gaming much since then.

I'm fine with repasting the cards - the laptop itself is what scares me, though. I have some Arctic Silver from a desktop I built a little while back - would that work, or is there something better?

 

Wow 104 is pretty high, did you try cleaning out the fans/heatsinks/heatfins? My 5870M CF while playing Dota2/SC2 doesn't usually go above 80's.

 

No, but it got up to 110°C earlier, still at <60% Load. I actually tried underclocking it, but the GPU Clock Tool will only affect the primary (setting it to the second one results in a failure, and it says to check the configuration...?).

I looked and found my Arctic Silver 5: since it seems like it really needs to be repasted, can I use it, or is the MX-4 much better?

 

I use the Arctic Silver 5 on my CPU and GPU heatsinks. It keeps them pretty cool (right now they're idling at 42C, and 49C). When cleaning your GPU heatsink, I find it worthwhile to run the fins under cold tap water for a bit. It removes any dust that films on the inside of the fins. That gets them much more cleaner than just blowing them with compressed air. Make sure to use the air can afterwards to get the heatsink completely dry. Also make certain to use a good cleaner to remove the old thermal paste from the heatsink and GPU die.

 

I disassembled it to clear out the dust (using Radji's method of using water on the fins, and letting them dry overnight). In the process of putting it back together, I killed the second GPU (ESD probably).

After repasting, the primary (and now only) GPU never goes above 88°C with Skyrim (which is surprisingly still playable!). I noticed that the sensor for the memory chips always reports significantly higher temps than the shader/other diode (by 10°C - 20°C). Is 88°C a safe operating temperature for the memory?

Also, I realized after breaking the second card that the higher temperatures can be mostly attributed to my having misinstalled the fans (pulling air away from the bottom, since I thought those vents were exhaust instead of intake). Convenient timing.

 

Sucks to hear I was going to ask if you had extra fans under the bottom vents or under the rear of the laptop where it is vented. Sounds like you found the problem as the bottom is obviously the intake

 

Yeah, I should've been able to figure it out based on the dust on the bottom, but it became much more obvious when I looked at the way the heatsink and fan were arranged!

 

Ouch. You've gotta be careful about wearing an ESD wristband or using an anti-static mat. FWIW, ESD might not have killed your 2nd card. You could have kinked the crossfire cable, or screwed the heatsink down too tight which damaged the die (which is pretty common). One point to note, computer fans always suck air in from the bottom, and blow it out the side or back. The only exception to this are Macs which use side by side vents.

 

I tried doing the retention clip mod, but I wasn't ever able to get any of the C-clips off. Is it possible to crack the die even with them in place? Also, I took the Crossfire cable out completely when I had tested it recently (e.g. had both cards in, but not connected); is there any way of knowing if I damaged the cable in the process?

For future reference, I want to clarify what I should do with anti-ESD stuff. If I have a wristband connecting me to the laptop, isn't the GPU still vulnerable if it's removed from contact with the laptop? Should I get an anti-static mat and use it in combination with my wristband (both connected to the chassis)? Also, are there any particularly good spots to latch the connectors onto? Every spot I tried the most recent time I had it open looked like it would've scratched the finish.

 

Here is some general overheating troubleshooting that you can use. Also try repasting as some other users recommend. It'll probably lower down the temps a bit.