I was just looking for some advice about whether I should do a warranty claim on my graphics card. I'm dithering because it's going to be a pain in the ass waiting a few weeks for it to be returned to me and I don't have a spare PCIE card so my computer will be offline during that time.
My card is a gigabyte 9600GT card which I bought in August this year. The problem is that I've been getting persistent artifacts while playing two different games (the witcher and fallout 3). The artifacts common to both games are red and blue hashing across the screen:
also in Fallout 3 there's somtimes this sort of SERIOUS warped pixelage:
These errors get slightly better if I turn anti-aliasing off but they do not go away, even on very minimal graphics settings.
The errors do only seem to creep in once the card warms up a bit. But I've used the Nvidia monitoring program to check the heat of the card and the highest I've ever seen it is 51 degrees which I'm pretty sure isn't too hot.
Strangely enough, I have played both Bioshock and Mass Effect on pretty high graphics settings with no errors whatsoever and good frame rates. Dunno what that means. I also can't create any errors by doing the artifacts test with ATI tools
I have of course updated to the latest nvidia drivers.
My question is, does this look like faulty manufacturing on the graphics card? My theory (and I don't know all that much about graphics cards so I could be way off) is that it's got a manufacturing defect and so its performance degrades at much lower temperatures than it properly should. But there's some possible alternatives:
* I put my rig together from scratch so I could have messed something up with the installation. I obviously can't see anything wrong, everything's plugged in correctly, the fan's spinning nicely, and there's no strange noises or anything. But is there something here that could be wrong?
* It's not my graphics card, its some other part of my machine (which was all bought at the same time as the card). Is this at all plausible?
idk, I can't think of any other explanations and I really don't want to be without a computer for a couple of weeks while the warranty goes through but I think I'm gonna have to bite the bullet.
Could be overheating caused by the heatsink not making proper contact with the GPU die, that can be quite a common thing with the cards produced in mass quantity, some manufacturers use the cheapest thermal compound and don't really care how well the contact is.
Try removing the heatsink and the old thermal compound and replacing it with a thin layer of Artic Silver 5, rubbing alcohol and some Q tips will remove the old compound, when the heatsink is removed you'll be able to see the old thermal coverage footprint, to see how well the GPU was covered in the first place, re-applying thermal compound will not mess up your warranty.
Case airflow is another factor, you can have a good GPU cooler but if the case airflow is choked out the fresh air cannot flow, and the case internal temp rises and causes GPU freezes and crashes, so airflow through your case is very important.
Artifacting is most prevalent in too high of overclocked cards, you haven't said whether you're overclocking the card, but if you are, the settings need to be backed off.