Hey, all. I recently made a post about some issues with my HD5970 locking up. After talking to a techinician from the company I purchased the computer from he came to the conclusion that it wasn't a driver malfunction, but simply that the card wasn't getting enough power and apparently ATI cards are very picky about power fluctuations. Now, while on the line I also brought up the fact that the card seemed to be running especially hot. He pointed me in the direction of a good temp gauge program (HWmonitor) and using that I read off the readings. At the time my video card was running @ about 53C. He told me that for idle that temperature was perfectly fine considering I'm using fan based cooling. He then went on to tell me what temps my computer should typically run at based on what I was doing. His quotes basically put me between 70 and 80 when gaming, and from 80-90 when I'm really pushing the card with intense 3D applications. He told me that if the card reached 90 or above that it was then at risk of burning out and I should definitely watch for that.
Well, the power issue was resolved almost instantly after speaking with him. I haven't had a lock up all day and the quality of the in game graphics has stayed consistent. My main concern is that I've been watching the reading for my video card temp and it's making me very nervous. When I initially started up the game it would fluctuate from any where between 72 and 80C, however when I started getting into areas or circumstances that would be graphically intensive I noticed the card climbing above 80. The highest it would spike to was about 85C before I would alt tab in which case it would cool back down to the high 70's - low 80's. I'm not entirely sure how high it would had climbed if I hadn't of backed off, but I didn't want to risk it. After a short while of it getting, from what I was told, dangerously hot, I went ahead and lowered the ingame draw distance and it cooled down considerably while in game. A good 5-7C to be exact.
I suppose my question is....is this irregular? Should I be concerned for my card or are there any steps I can take to make it run cooler? By his numbers I would be led to believe that the max. stable running temp for a HD5970 is 90C. Is that right? Just as a side note, the game in question is World of Warcraft and I'm running it at absolute max settings in 1680x1050 resolution. I never thought of WoW as a graphically demanding game, but after running it at ultra settings it's like an entirely new game. After spending so much money it would seem ludacris to me that I couldn't run this generation's games @ max settings without worrying about my video card melting. Any help or advice is greatly appreciated.
If you're running any hotter than that (about mid 90's) then there might be a problem. That doesn't necessarily mean that your card is in danger of burning out since many high-end GPUs are able to run very hot, but you should wonder if something is wrong with the card's cooling. Excessive dust is the main culprit of overheating in computer components. A loose heatsink can also be an issue, but I'm certain that is not an issue you'll have any time soon. With that said, if your card runs up to 85C, there is nothing wrong with it.
Thanks to the both of you for your help and information. It will definitely put my mind at ease. Honestly I assumed with such precision technology a heat so high would be risky, but I take it the hardware is extremely durable. My only concern now was brought up from your link, skolpo. I can see that my load temp is right around where it should be, but why is my idle temp so high in comparison to those readings? Perhaps I'm misunderstanding what idle actually means. Is that while your computer is in sleep mode or something to that degree? Because the 53C reading I'm getting is while I am, for example, on this website replying to this post. I suppose idle means while it's doing nothing at all? Because even looking at my desktop requires a bit of graphical proccessing power.
I'm also curious as to if the 85C reading was my card at it's worst. I had stepped away from my computer while in transit ingame (flightpath in World of Warcraft). Such a thing is fairly demanding graphically since you traverse the terrain rather quickly and the draw distance is quite far with quite a bit of detail. I'm hoping it peaked at 85C during this, and was regulated by the fans as it should be. When I came back I had just landed at my destination and the card had already dropped to ~81C. I honestly believe I'm just being paranoid, but I'd very much like to protect my investment!
My next order of business will be to update the drivers, something I'm completely in the dark about. As an example I had recently tried to install them and was unaware you had to uninstall the former drivers prior installing the new ones. As a result my monitor would go black about midway through installation. I hope this is attributed to my blunder, and not some legitament issue with the card itself. I'm also hoping that trying to double dip on the drivers hasn't somehow effected it's performance or messed with it's current drivers somehow. Should I expect a performance increase and/or better run temps with more recent drivers? Perhaps with tweaks made to automatic fan speeds while under load? Again, I really do appreciate the help. I don't know nearly as much about computers as I'd like to, : .
Also, one last thing I'd like to add. I'm running with a stock i7 920 @ 2.67. Since I've been looking at HWmonitor I've noticed my CPU core temps as well. So far the readings I have been getting are 44-46C lows and 61-63 highs. The CPU is liquid cooled via Alienware cooling. Just curious how those temps sound. The highs are most likely from when I was playing World of Warcraft @ Ultra settings with other apps still up out of game as well.
Anand's idle temp shows when the card goes into 2D mode (frequencies are clocked down drastically) while regularly it will idle much higher than that. It might also be a fluke reading since I can't find another review that drew the same temperatures at idle. Several other reviews show the idle temp around 55C-60C and the load goes up to 87C. During display driver installations, your monitor will flash and go dark. That is also normal. You can go ahead and relieve all of your paranoia.
Alright, well again thank you all very much! It appears my system is in tip top shape, . The paranoia has finally subsided and I can once again use my computer with confidence, . You have all been extremely helpful and I honestly can't thank you enough.
Also I'd like to note that last night I left my computer on in sleep mode. When I awoke I checked HWmonitor and actually had a low reading of 34C. So it's possible that's where they got their idle temps from. If that's the case ~34 - ~54 - ~85 seem to be aceptable temps at idle - 2D - 3D.
I am running 2 2GB Sapphire 5970's at default frequency in crossfire mode. I played Crysis at max settings last night and my PC shot off after about 12 minutes of doing that. I hadn't been monitoring the video card heat but they nearly burned my skin when I reached around to feel them after the crash. That really caused concern so I booted the system and ran the fans at 100% until the heat got back down to under 60 degrees. I wanted to monitor the cards today so I could see how hot they were actually getting with the ATI CCC temperature gauge. I turned the fans on full blast while I played Crysis (not in full screen this time) and the heat this time never got over 86 degrees even at 98% card usage... but it did get to 86 nonetheless. The system just needed more airflow. I'll probably get a gpu specific cooling solution before too long but at this point the cards are at least running at an acceptable level, even if just barely.