I need some help and hope this is the right section, I couldn't find a specific support forum and because I think the GPU is my problem I put this here. Quick system overview:
i7 920 D0
6GB Corsair Dominator
650W Corsair PSU
128GB Crucial M4
640GB Western Digital
Not used my comp for a few months (summer), came back to uni and it worked fine for 4 heavy days. Then I moved it to another room, didn't use for a day or two, then when I did I noticed it was being a pain with the monitor not picking up signal (HDMI) so I tried DVI and it was fine. Using the PC, a couple times the display driver crashed and instantly recovered on its own, but every 5 mins or so I'd get weird artefacts all over the screen.
This soon turned into the system not booting into Windows, and the boot screens would have weird glitchy artefacts and stuff on screen, from thick lines vertical then weird green stripes filled with 9s (9999999 etc) and then triangles. it let me boot into safe mode but then it would have angled lines running down the screen.
Assuming this was my GPU I requested an RMA (I bought it 28th Nov last year... after I killed my 5750 by spilling drink and splashing the case somehow.
After a few days (now) I decided I'd give it one last shot of switching around the PCI slot and unplugging all the cables to the card (again). Same again [artefacts] except this time it said I'd installed a new CPU and it couldn't continue... (might have something to do with PC case being on its side) so I rebooted a few times and kept getting the same as before; artefacts, black screen past Windows logo with no boot sounds. THEN, I tried again and this time... no artefacts, boots straight into Windows and here I am typing this message... WTF?
I'll upload video shortly. Any advice on the issue? I don't want to have to RMA my card cause I'd rather save the hassle and not be trying to do heavy design uni work on an ultrabook... ESPECIALLY if it turns out not to be the card and they return it too me! But at the same time I don't want to just drag it out now while it appears to be fine, only for it to die 6 weeks down the road when I'm out of my 1 year warranty.
Blow all dust out of all the heatsinks and exhaust/intake ports on the computer. Use temp monitoring software like CPUID HWMonitor or CoreTemp and GPU-Z. If you can't get into the OS most BIOS have a temperature monitoring screen (although it won't show you GPU temp). Report your temps if you aren't sure how to interpret them.
Aside from heat, your problem sounds a lot like a dying GPU. But it could be many things. If it isn't heat, and you've made sure your drivers are up-to-date then find a known-working system (that can handle a 6870) and test your 6870 in it. I'd start by first eliminating heat and drivers as an issue and then testing the GPU in a different system.
It's possible it's a bad video card, there's also a chance it's bad RAM. So, the first thing to do would be to get a copy of Memtest86+ and let that run overnight. If you're not showing any errors, then I'd say it's worth giving the video card RMA a shot.
Thanks for the reply guys but while I was using the comp, it (2 or 3 times) gave me frantic checkboard patterns down each side of the screen before it'd free, driver crash and recover, and all is well again for 5 minutes. (drivers were updated a week or two ago)
I tried putting something strenuous on the card so I can furmark burn in, assuming it'd crash instantly but it actually held up well for the 5 minutes I ran it. Fan quickly went into full spin but it started the temps around 40 something and raised them to 76 by the time I just quit out. THEN not long after, the same checkerboard thing happened, but 2-3 times after the other until I had to reset the PC, which got to where the W7 login screen should be (boot audio) but just displayed a dull blue colour and didn't shift.
I may still try and get into safe mode if I can and run Memtest but to me the more it's looking like the gpu! : \