Here's something I posted to the Geforce boards as well, when I thought I was having a driver issue:
 
All of a sudden, my PC (Intel i7 920, GTX 570 SC HD, Windows 7 64 bit) is unable to play any games, or even use programs like 3DMark 11. I mainly play Star Trek Online, Team Fortress 2, Skyrim lately, but now I'm at the point where basically no games work except those that don't "Stress" the system.  Nothing is overclocked except the card's factory overclock.

What happens is, in all cases, the game(s) will let me past the loading screens and whatnot, and even occasionally let me play the game for a few seconds.... then the screen goes black then solid color light green, light blue, greenish blue, etc.... requiring a reset or cold shutdown. One time I did get a blue screen mentioning nvlddmkm.sys, failed to respond in a timely manner, etc... but I can't seem to pull up that exact event in Event viewer.

Anyway, when this started happening, it was "only" happening in Skyrim - I started STO and it ran fine for a minute or so. I wondered if it might be an OpenGL issue, so I went into Starbound - which has "normal" and OpenGL modes - Starbound would work in "normal mode", but it did the crash if I launched it in OpenGL mode.

I did notice on the forced reboots that I keep getting a message the it was "Updating Steam", so I wondered if it was a Steam thing. I uninstalled 340.43 and let the pc restart, then reinstalled it.

What that did, is make it so that NO "accelerated app" would work - not even Starbound or STO. That's the funny thing... it "waits" until I actually would get into the game.

I don't think it's a heat issue - This room is small and air-conditioned, with the A/C less that four feet away. I even opened up the PC and checked for dust - there wasn't any - I'd cleaned it out a few months ago with canned air. I did remove the vid card, and blow canned air into it, and into the case, just to be sure. The PSU is a Corsair TX650W, and that didn't have much dust on it either, though I did blow some air through it, too.  There was far, far more dust in the system the last time I cleaned it out two months ago, and it was working fine then.

Any ideas or suggestions? Aside from trying to play games, the PC is working; I'm using it now. EGA Precision X shows proper temps and fan speeds... and the test function works.  Those brief moments I am in a game, there isn't any image corruption - this card is an RMA from another GTX570 SC HD that died - that scrambled the screen everywhere.

So, I'm thinking some sort of software issue, but darned if I can find it.

Any ideas? I did a Norton Quickscan right now, and I'd done a Spybot 2 scan last night, all clear.

Any other info from my PC that might help?
 
The thing that gets me is it was basically restricted to OpenGL games until I reinstalled the drivers the first time.  After that, nothing works.  Well, except Civ 4, which isn't exactly a system killer. The system passes both 32m and 1024m 8-threaded wPrime tests, and will run the EVGA OC scanner stress test at 95%+ load just fine with zero artifacts.

ty_ger07
What is the video card core temperature while OC Scanner is running?

Usually a solid color screen crash is related to a power issue. Either a PSU issue, an issue with the video card making bad connection with the PCI-E slot, an issue with the motherboard's 24-pin connection, an issue with the video card's PCI-E 6-pin connection at the video card or at the PSU (if modular PSU connection design), or an issue with the video card's power regulation ("VRM") circuitry.

Okay, that was weird...
 
I just ran the OC scanner stress test again.... this time while watching Precision.  When it started, the clock speed bounced between 797 mhz (the card's factory stock OC'ed speed) and 50 percent (398), but it settled in at 398 mhz while the temp went to 72c, while at 96 percent load.
 
I stopped the test, and noticed that the artifact scanner was off.  I turned it on, then tried the stress test again - got the green screen crash instantly.
 
Boy, this bugs me.  I still have no idea if it's a PSU, GPU, or even a software issue.  I'd try my GTX 275 CO-OP, but it's debatable if that has higher or lower power requirements than the 570.
EDIT: I did unslot and reslot the card yesterday when I was cleaning it out, but I didn't do anything with the power cables - they seemed to be in place.  Tried unplugging and replugging the card and the power cables, blew out a tiny bit more dust, no good.
 

MightionNY I would recommend going through with the RMA as long as your warranty is still active and if you are interested in upgrading to another GPU you could always sell the replacement and use that towards a new purchase.  
 
Please email us at Support@EVGA.com when you are able to and we can help you out.

MightionNY
Well, according to the details of this card, it's out of warranty; even though the first card I bought new had limited lifetime warranty.  I had to RMA that one, for this one.

Please contact us by either starting a support request from your "My Products" page, or contact us via phone at 1.888.881.3842, and we can assist you in setting up your RMA.