Hi, I'm hoping you can help me out. I have a PowerColor HD 6950 Vortex II 2 GB in my machine. It keeps shutting itself off at 50c. I'm running a cooler master silent pro gold 1000w psu so I don't think I have a problem with insufficient power. I'm using the HDMI and using a 42" HDTV for a screen, running resolution of 1920x1080, it is a 1080 tv. I've tried completely uninstalling and reinstalling my drivers. I've tried taking it out and fully cleaning it. I've tried applying new thermal paste(Antec formula 7). It does seem to have the fan speed bug where it won't ramp up the fan automatically as temps increase. I've tried using CCC to manually ramp up the fan to 95% however that just delays how long it takes for the temp to hit 50c and the card shutdown. Any ideas or suggestions you can give me I would be appreciative of. If I need to provide more information I'll also be happy to do so.
And do you get any errors, or info popups or anything? Kinda hard to work with just 'shutting down' It 100% turns the display off, or the driver crashes and recovers?
Driver crashes, does not recover. It just goes black screen as soon as it hits 50c. Audio keeps playing, and the rest of the machine keeps running. Just the display stops working. It's almost as if it is overheating and shutting down to prevent damage, but 50c is well within tolerances. If you have a recommendation for test software to output an error log I'd be happy to run it and post the results, but I'm not sure what to run to generate a log.
Is the GPU overclocked? If you haven't already, use DDU to remove the old drivers in safemode http://www.wagnardmobile.com/DDU/ And Event viewer is gonna be your best bet for finding out what's going on, although I imagine it will just say display driver stopped responding etc etc GPUs throttle to cure overheating, they don't crash the driver, it's not overheating anyway, 50c is fine
GPU is not overclocked. When I use DDU to remove the old drivers in safemode, should I first uninstall the new drivers so I'm basically going to a clean slate and then re-install the new drivers?
Nah, DDU does everything for you - just run it from normal mode, it will boot you into safemode, then press the recommended box and sit back
Just got done running DDU and then installing the latest stable AMD drivers available. About to run Heaven benchmark and see what happens. EDIT: Display crashed again. I'm thinking it is either not a driver issue or I should try something other than the latest stable drivers. Do you think I should try the beta driver and see if that makes a difference?
Might be your memory modules or some other non-GPU component on the PCB are stressed, cooling not applied sufficiently. Did you try that stupid 6970 shader unlock by modifying the bios, if yes then kiss your card goodbye.
May be caused not by GPU issue, but VRMs issue. But usually if card for any reason shuts down PC it is full emergency shutdown = no power. My guess is driver incompatibility, Does it happen in OCCT burn test which is windowed? Does it happen even under very low loads like GPUz PCIe test?
My cards run up to 100% on stock coolers with 6790 unlocked bios with no issues. Either your heatsink has a problem or the gpu temp sensor is off and is broken. I am not sure how you can fix the internal temp sensor. Double check your case air flow and psu air flow. If all else fails, try another gpu to make sure the rest of your system is still working with out problems and buy a new GPU.
I'd suggest manually setting the Fan speed to a constant 50% and checking it once, depending upon your regional ambient temperature. 45-50%, try it and check if the card is stable.
I'm not running the 6970 shader unlock, just the default bios. If it is another component of the card that isn't cooling sufficiently, what would be an appropriate fix? It does happen with the GPUz PCIe test with the test windowed. I have not tried the OCCT burn test. If it's a voltage issue, does that mean my power supply might be going bad? Is there an easy way to check that? If it is a driver incompatibility, should I try the beta drivers or find an older stable version? EDIT: I hope it's not a physical defect with the card, if it is, then it is. The only other GPU I have lying around is an old Nvidia 9600 GT, would that work for testing?
if you can glue heatsink on it it may work. No, I would use DDU to clean everything up (how to is on forums here), then instal WHQL which is not that different from beta. Yes, but before you remove 6950 use DDU to remove all AMD drivers. Then shut down, place in 9600 GT and start again.
Definitely try that 9600 GT to make sure your system is ok. How do I do this with CCC? I set and forget my settings and yes, 100% fan CFX is freaking loud!
it's through the Overdrive tab in CCC. as a fellow 6900 user I've had bad results (not instability but I didn't keep it long enough to test) with the latest WHQLs and most 14.x drivers, until the recent 14.11 betas that seem to work well, so I wouldn't rule out drivers just yet. and yes your 9600 GT should do the trick to see if it's a GPU issue or some other hardware acting up.
Not sure on the temp but I'm getting the exact same problem. This problem has only just started occurring this month. All running fine until recently and for a long long time. I have both a 6950 and a 6970 which were xfired. Full uninstall of all drivers with ddu and tested with old nvidea 8600gts in sli and all is fine. It just seems as though the cards switch off. I have tested both independently and both have same issue. I have a Asus P8P67Pro. Have updated the bios, the drivers. Have run intel cpu testing tool. All fine. I suspect driver but have tried older drivers and they did not make the problem go away. Have not tried beta driver yet. Is it possible that new 14.x driver updated the cards bios?