So hey guys, I had this problem for quite a while, and I'd like your opinion on it. I bought this card (ASUS Radeon HD 7970 Directcu II) about 2 years ago. >tinypic [dot] com/r/2ptyyvq/5 For about the first year, everything was perfect, would run all games on ultra etc. After some time, artifacts started appearing, and then my games would crush in the high settings with stripped lines appearing everywhere or my screen would just go black (and the sound would **** up as well). Only think I could do was restart. Even if I tried the games in windowed mode, the game would stop responding, and a pop-up would tell me that the Graphics Adapter has stopped working but has recovered. Oh, and to get that out of the way, I always run the latest drivers (updated from catalyst 13.9 to 13.11 a couple of days ago). So, I know that the 7970 is actually an overclocked version of the 7950. So I got myself a copy of MSI Afterburner and used it to increase the voltage, little by little, until my card eventually stabilized. With the increased voltage, my card would play everything in high/ultra again with no problems. But recently, I got BattleFeel 4 and problems started again.. Now, no matter how much I change the voltage and the fan speed, the game will go for a minute or two before crushing and forcing me to restart the machine. i39.tinypic [dot] com/30biidh.jpg Could it be a thermal issue? I mean, the only reason I bought this massive GPU is because it looked like it came with a decent cooler and had good airflow (my previous gpu, the 6970, cooked itself). I was wondering if any other people here have this problem with this card, or if they know of an optimal setting or some solution to this. I usually run my card with these afterburner settings: i43.tinypic [dot] com/awyo1v.png Kudos! Update: Ok, I played some test range with BF4 in windowed mode while monitoring the temperature. It climbs up from 25 degrees Celsius at about 1 degree per second, at about 45 degrees, artifacts start to appear, at 55 degrees, severe artifacts and at 65 degrees, the game crashes and stops responding. Should I remove the heatsink and reapply thermal paste on the chips? I mean, my previous card, the 6970 cooked itself because I didn't apply enough paste at one of the chips, so it could be dangerous.. but if it's going to make my card work properly again.
65°C is not temperature at which any GPU would start artifacting. If anything is overheated then it's memory. If you were not running card over 1000MHz at core and 1400MHz for memories, then there is no reason for damage caused by high OC. 1.15V is not anything too high either, load GPUz, click top left corner and "Read ASIC Quality". Post it here so we know what approximately should be your default voltage. You may actually need to raise memory voltage instead of core. BTW, 7950 is not downclocked 7970. 7950 is cut down and downclocked. One possibility which should not be overlooked can be unstable PSU or bad contact on PSU connector for card which may lead to insufficient current and high Vdrop.
ASIC quality is 82.8% I managed to stabilize the GPU for now with 1050 mV, 1000 MHz as Core Clock and 1200 MHZ as memory clock. Played one full game with no artifacts or crashes at high settings.
Mine is 81.5% and default voltage under load is 1.09V moved it to 1.14V to get OC from 1050MHz core to 1125MHz. I see that you kept core, but reduced memory clocks. Your DCII cooling include passive block for memories, but 3 of them are covered only by 1/2 which may cause overheating. Cooling for GPU itself is strong even in case you have bad airflow in PC, but memories are cooled same way as VRMs by front cover plate which needs proper airflow. Can you read VRMs temperature under load by GPUz/HWiNFO64?
