I've been having some strange issues with my XFX Radeon HD 6770 recently. I've been using GPU-Z to log the various sensors and I'm finding that it rarely goes over 50% load, peaking at around 56% occasionally, yet it reaches 70+*C and doesn't turn the fan up any higher than 50%. I've been playing Sims 3 a lot with everything maxed out and after about 20 minutes when the GPU gets warm the screen flickers and sometimes goes completely blank, leaving the computer stuck and needing a hard reset. Occasionally I can alt-tab back to the desktop where the task bar flickers for a bit, but after 5-6s it's all cooled down and I can go back into the game.
I'm using the built in AMD Catalyst software with the card and have it overclocked via AMD OverDrive to 950MHz GPU clock, 1375MHz memory clock, where it has been running 100% stable for several months until this apparent temperature issue. Is there a way I can adjust the fan speed so that it kicks up a gear when the temperature reaches say 65*C instead of floating around 50% at 70*C, also is there any reason why the GPU load only ever goes up to around 50%? I'm not just talking about in Sims 3, I find the framerate gets a bit laggy running games like Prototype or Dead Island at higher settings, but the GPU still isn't working to it's full potential.
System specs are:
Asus P5Q-E Motherboard
Core 2 Duo E4500 CPU (OC: 333MHz FSB, x9 multiplier, 3.0GHz core, 1.320V - Prime95 tested for 24H and stable below 70*C on stock cooling)
8Gb (4 x 2) OCZ Titanium DDR2-800 @ 400MHz (4-4-4-15 2.2V)
XFX Radeon HD 6770 (OC: 950MHz GPU, 1375MHz memory)
OCZ Stealth Xstream 2 - 600W PSU
Not sure what my case is but I've put an additional Arctic Cooling F12 120mm exhaust fan on the back as well as fully sealing the case with dust & dirt filters, so it's nice and clean inside with plenty of air flowing through.
Ok I will link you to MSI Afterburner you can monitor your temps while you play your game and it is on the screen the whole time while playing. You can set way points for your fan to in crease as the temps do.
Also displayed is your fps and % usage on your GPU. And if you know anything about OC'ing you can use that program to OC your card. It is a nice little program and I use it all the time.
One more thing if your temps go over 80c you will start to have problems like lagging, crash to desktop (CTD) screen flicker etc. Even if your cards shows it can go to 100c. What that means when it gets there no more card.
This program is free plus I will link you to the manual and to another burn in test and a few more temp monitoring programs for the coer and CPU. I hope this helps and good luck to you.
Thanks for those. I've had a play using Afterburner & combustor. Seems that temperature actually wasn't the problem and that lowering the memory speed slightly solved it, but either way I've programmed the fan speed to hit 100% at 80*C - it's not a top-of-the-range card and the pc spec is generally a bit dated now so there's nothing to gain from pushing to the absolute limit, I'd rather the reliability.
It's strange because it had been running fine at that speed for a few months and never overheated nor showed any other signs of trouble. Perhaps the card is generally a poor design and is slowly failing anyway. I wouldn't be entirely surprised if that were the case - this is the 2nd one of these cards I've had as the first one was defective from the factory.
In fact I've now found that using Afterburner I can overclock it further to the maximum 960MHz core clock and 1380MHz memory, and it's more stable like this! Is it normal that certain combinations of core/memory clock might just naturally lend themselves to being less stable than higher frequency combinations?
You never know it could be that is what your card needed to unlock it to play better and some of the 6xxx series cards OC very well and are very stable. Now your card was before reaching high temps.
Do to your fan not working the right way and now your fan zooms along keeping your temps and lower temps. If your problems have gone down some then you know you are on the right track.
Card get old and heat is any electronic's worst enemy. All cards should be rated on hours and not years. Because if you game everyday for at least 4 to 6 hours a day. That card will wear out fast.
So your card could have been slowly dieing and with a little OC it's back in the game. Just watch your temps all the time and your CPU temps as well. I hope it all works out and let us know if it doesn't. Good luck to you.